AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 AI Matters

Annotated Table of Contents AI Ethics Education

Welcome to AI Matters 4(4) AI Policy Matters Amy McGovern, co-editor & Iolanda Larry Medsker Leite, co-editor Full article: http://doi.acm.org/10.1145/3299758.3299765 Full article: http://doi.acm.org/10.1145/3299758.3299759 Welcome and summary TAI Policy Matters

Call for Proposals: Artificial Intelli- Epochs of an AI Cosmology gence Activities Fund Cameron Hughes & Tracey Hughes Sven Koenig, Sanmay Das, Rosemary Full article: http://doi.acm.org/10.1145/3299758.3299868 Paradis, Michael Rovatsos & Nicholas AI Cosmology Mattei Full article: http://doi.acm.org/10.1145/3299758.3299760 Artificial Intelligence and Jobs of the Artificial Intelligence Activities Fund Future: Adaptability Is Key for Hu- man Evolution AI Profiles: An Interview with Iolanda Sudarshan Sreeram Leite Full article: http://doi.acm.org/10.1145/3299758.3300060 Marion Neumann AI and Jobs of the Future Full article: http://doi.acm.org/10.1145/3299758.3299761 Interview with Iolanda Leite Links

Events SIGAI website: http://sigai.acm.org/ Michael Rovatsos Newsletter: http://sigai.acm.org/aimatters/ Blog: http://sigai.acm.org/ai-matters/ Full article: http://doi.acm.org/10.1145/3299758.3299762 Twitter: http://twitter.com/acm sigai/ Upcoming AI events Edition DOI: 10.1145/3284751

Conference Reports Join SIGAI Michael Rovatsos Full article: http://doi.acm.org/10.1145/3299758.3299763 Students $11, others $25 For details, see http://sigai.acm.org/ Conference Reports Benefits: regular, student Also consider joining ACM. AI Education Matters: A Modular Ap- proach to AI Ethics Education Our mailing list is open to all. Heidi Furey & Fred Martin Full article: http://doi.acm.org/10.1145/3299758.3299764

1 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 Notice to Contributing Authors to SIG Newsletters Contents Legend

By submitting your article for distribution in this Book Announcement Special Interest Group publication, you hereby grant to ACM the following non-exclusive, per- Ph.D. Dissertation Briefing petual, worldwide rights: • to publish in print on condition of acceptance AI Education by the editor • to digitize and post your article in the elec- Event Report tronic version of this publication Hot Topics • to include the article in the ACM Digital Li- brary and in any Digital Library related ser- vices Humor • to allow users to make a personal copy of AI Impact the article for noncommercial, educational or research purposes AI News However, as a contributing author, you retain copyright to your article and ACM will refer Opinion requests for republication directly to you. Paper Precis´

Submit to AI Matters! Spotlight We’re accepting articles and announce- ments now for the next issue. Details on Video or Image the submission process are available at Details at http://sigai.acm.org/aimatters http://sigai.acm.org/aimatters.

AI Matters Editorial Board

Amy McGovern, Co-Editor, U. Oklahoma Iolanda Leite, Co-Editor, KTH Sanmay Das, Washington Univ. in Saint Louis Alexei Efros, Univ. of CA Berkeley Susan L. Epstein, The City Univ. of NY Yolanda Gil, ISI/Univ. of Southern California Doug Lange, U.S. Navy Kiri Wagstaff, JPL/Caltech Xiaojin (Jerry) Zhu, Univ. of WI Madison Contact us: [email protected]

2 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Welcome to AI Matters 4(4) Amy McGovern, co-editor (University of Oklahoma; [email protected]) Iolanda Leite, co-editor (Royal Institute of Technology (KTH); [email protected]) DOI: 10.1145/3299758.3299759

Welcome to our new co-editor! and September 30, 2019. Nominations are welcome from any country, but only English Welcome to the fourth issue of the fourth vol- language versions will be accepted. Only one ume of the AI Matters Newsletter! Our biggest nomination may be submitted per Ph.D. grant- news is that we would like to extend a big wel- ing institution, including large universities. Dis- come to our new co-editor! Iolanda Leite is sertations will be reviewed for relevance to ar- an Assistant Professor at the School of Com- tificial intelligence, technical depth and signif- puter Science and Electrical Engineering at icance of the research contribution, potential the KTH Royal Institute of Technology in Swe- impact on theory and practice, and quality of den. Iolanda is excited to join the AI Mat- presentation. The details of the nomination ters team and to help to continue to grow the process will be announced in early 2019. newsletter for ACM SIGAI. She is also our guest for the interview this month so please go to our interview and learn more about our new co-editor! ACM SIGAI Student Essay Contest The ACM Special Interest Group on Artificial Announcements Intelligence (ACM SIGAI) supports the devel- opment and responsible application of Artifi- Joint AAAI/ACM SIGAI Doctoral cial Intelligence (AI) technologies. We there- Dissertation Award fore invite essay submissions to the 2018 The Special Interest Group on Artificial In- ACM SIGAI Student Essay Contest, to be pub- telligence of the Association for Computing lished in the ACM SIGAI newsletter AI Mat- Machinery (ACM SIGAI) and the Association ters, addressing one or both of the following for the Advancement of Artificial Intelligence topic areas: (AAAI) are happy to announce that they have established the Joint AAAI/ACM SIGAI Doc- • What requirements, if any, should be im- toral Dissertation Award to recognize and en- posed on AI systems and technology when courage superior research and writing by doc- interacting with humans who may or may not toral candidates in artificial intelligence. This know that they are interacting with a ma- annual award is presented at the AAAI Con- chine? ference on Artificial Intelligence in the form of a certificate and is accompanied by the op- • What requirements, if any, should be im- tion to present the dissertation at the AAAI posed on AI systems and technology when conference as well as to submit one 6-page making decisions that directly affect hu- summary for both the AAAI proceedings and mans? the newsletter of ACM SIGAI. Up to two Hon- orable Mentions may also be awarded, also Win one of several $500 monetary prizes or with the option to present their dissertations a Skype conversation with a leading AI re- at the AAAI conference as well as submit one searcher, including Joanna Bryson, Murray 6-page summary for both the AAAI proceed- Campbell, Eric Horvitz, Peter Norvig, Iyad ings and the newsletter of ACM SIGAI. The Rahwan, Francesca Rossi, or Toby Walsh. award will be presented for the first time at the AAAI conference in 2020 for dissertations that The deadline for submissions is January have been successfully defended (but not nec- 10th, 2019. essarily finalized) between October 1, 2018 For details, please visit the blog announce- Copyright c 2018 by the author(s). ment.

3 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

*New Conflict of Interest and Submit to AI Matters! Retractions Policies Thanks for reading! Don’t forget to send The ACM Publications Board has approved your ideas and future submissions to AI new Conflict of Interest and Retractions poli- Matters! We’re accepting articles and an- cies. Both of these new policies are now nouncements now for the next issue. De- posted on ACM’s website at: tails on the submission process are avail- https://www.acm.org/publications/policies/conflict- able at http://sigai.acm.org/aimatters. of-interest

Issue overview This issue opens with an interview from Mar- ion Newmann of our new co-editor: Iolanda Leite. We also have a summary of recent AI conferences and events from Michael Rovat- sos. Our educational column this issue pro- vides a model AI assignment for teaching ethics in AI focused on autonomous vehicles. Amy McGovern is co- Larry Medsker’s policy column summarizes a editor of AI Matters. She variety of policy issues of interest to the SIGAI is a Professor of com- community including the creation of a new puter science at the Uni- Roadmap for Artificial Intelligence, FTC hear- versity of Oklahoma and ings on AI, and recent discussions on how to an adjunct Professor of design AI to be safe as well as issues sur- meteorology. She directs rounding legality of AI. He closes with a short the Interaction, Discovery, discussion of privacy for AI Matters. ACM’s Exploration and Adapta- agreement with EasyChair specifies that we tion (IDEA) lab. Her re- collect age and gender data and we welcome search focuses on ma- any discussion on our proposed response to chine learning and data mining with applica- this data collection. tions to high-impact weather. Finally, we close with two paper contributions. Iolanda Leite is co-editor The first is a continuation of the discussion of of AI Matters. She is an an AI “cosmology” by our contributing editors Assistant Professor at the Cameron Hughes and Tracey Hughes. They School of Computer Sci- propose the idea of epochs of AI in their cur- ence and Electrical Engi- rent article. The second paper discusses the neering at the KTH Royal future of AI and how it will affect jobs. Institute of Technology in Sweden. Her research in- terests are in the areas of Human-Robot Interaction and Artificial Intelli- gence. She aims to develop autonomous so- cially intelligent robots that can assist people over long periods of time.

4 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Call for Proposals: Artificial Intelligence Activities Fund Sven Koenig (elected; ACM SIGAI Chair) Sanmay Das (elected; ACM SIGAI Vice-Chair) Rosemary Paradis (elected; ACM SIGAI Secretary/Treasurer) Michael Rovatsos (appointed; ACM SIGAI Conference Coordination Officer) Nicholas Mattei (appointed; ACM SIGAI AI and Society Officer) DOI: 10.1145/3299758.3299760

ACM SIGAI invites funding proposals for arti- mation on at most 3 pages: ficial intelligence (AI) activities with a strong outreach component to either students, re- • a description of the activity (including when searchers, or practitioners not working on AI and where it will be held); a budget for technologies or to the public in general. the activity and the amount of funding re- quested, and whether other organizations The purpose of this call is to promote a better have been or will be approached for fund- understanding of current AI technologies, in- ing (and, if so, for how much); an explana- cluding their strengths and limitations, as well tion of how the activity fits this call (including as their promise for the future. Examples of whether it is new or recurring, which audi- fundable activities include (but are not limited ence it will benefit, and how large the audi- to) AI technology exhibits or exhibitions, hold- ence is); ing meetings with panels on AI technology (in- cluding on AI ethics) with expert speakers, • a description of the organizers and other creating podcasts or short films on AI tech- participants (such as speakers) involved in nologies that are accessible to the public, and the activity (including their expertise and holding AI programming competitions. ACM their affiliation with ACM SIGAI or ACM); SIGAI will look for evidence that the informa- • a description of what will happen to the sur- tion presented by the activity will be of high plus in case there is, unexpectedly, one; and quality, accurate, unbiased (for example, not • the name, affiliation, and contact details (in- influenced by company interests), and at the cluding postal and email address, phone right level for the intended audience. number, and URL) of the corresponding or- ACM SIGAI has set aside $10,000 to provide ganizer. grants of up to $2,000 each, with priority given to a) proposals from ACM affiliated organiza- Grantees are required to submit a report to tions other than conferences (such as ACM ACM SIGAI following completion of their ac- SIGAI chapters or ACM chapters), b) out-of- tivities with details on how they utilized the the-box ideas, c) new activities (rather than ex- funds and other information which might also isting and recurring activities), d) activities with be published in the ACM SIGAI newsletter “AI long-term impact, e) activities that reach many Matters”. people, and f) activities co-funded by others. The deadline for submissions is 11:59pm on We prefer not to fund activities for which suf- March 15, 2019 (UTC-12). Proposals should ficient funding is already available from else- be submitted as pdf documents in any style at where or that result in profit for the organiz- ers. Note that expert talks on AI technology https://easychair.org/ can typically be arranged with financial sup- conferences/?conf=sigaiaaf2019. port of the ACM Distinguished Speaker pro- The funding decisions of ACM SIGAI are fi- gram (https://speakers.acm.org/) and nal and cannot be appealed. Some fund- then are not appropriate for funding via this ing earmarked for this call might not be call. awarded at the discretion of ACM SIGAI, A proposal should contain the following infor- for example, in case the number of high- quality proposals is not sufficiently large. In Copyright c 2018 by the author(s). case of questions, please first check the

5 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

ACM SIGAI blog for announcements and clarifications: https://sigai.acm.org/ aimatters/blog/. Questions should be di- rected to Sven Koenig ([email protected]).

6 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

AI Profiles: An Interview with Iolanda Leite Marion Neumann (Washington University in St. Louis; [email protected]) DOI: 10.1145/3299758.3299761

Introduction Getting to Know Iolanda Leite This column is the seventh in our series pro- When and how did you become interested filing senior AI researchers. This month we in CS and AI? are happy to interview Iolanda Leite, Assistant I became interested in CS at the age of 4 when Professor at the School of Computer Science the first computer arrived at our home. It is and Electrical Engineering at the KTH Royal more difficult to establish a time to define my Institute of Technology in Sweden. This is a interest in AI. I was born in the 80s and have great opportunity to get to know Iolanda, the always been fascinated by toys that had some new AI Matters co-editor in chief. Welcome on level of “intelligence” or “life-likeness” like the board! Tamagotchi or the Furby robots. During my Master’s degree, I chose the Intelligent Sys- tems specialization. That time was probably when I seriously considered a research career in this area.

What professional achievement are you most proud of? Seeing my students accomplish great things on their own.

What would you have chosen as your career if you hadn’t gone into CS? Figure 1: Iolanda Leite I always loved to work with children so maybe something related to child education. Biography What do you wish you had known as a Iolanda Leite is an Assistant Professor at Ph.D. student or early researcher? the School of Computer Science and Electri- cal Engineering at the KTH Royal Institute of As an early researcher I often had a hard time Technology in Sweden. She holds a PhD in dealing with the rejection of papers, applica- Information Systems and Computer Engineer- tions, etc. What I wish the “past me” could ing from IST, University of Lisbon. Prior to join- know is that if one keeps working hard, things ing KTH, she was a Research Assistant at the will eventually work out well in the end. In Intelligent Agents and Synthetic Characters other words, keeping faith in the system. Group at INESC-ID Lisbon, a Postdoctoral As- sociate at the Yale Social Robotics Lab and an What is the most interesting project you Associate Research Scientist at Disney Re- are currently involved with? search Pittsburgh. Iolanda’s research inter- ests are in the areas of Human-Robot Inter- All of them! If I have to highlight one, we action and Artificial Intelligence. She aims to are working with elementary schools that have develop autonomous socially intelligent robots classes of newly arrived children in a project that can assist people over long periods of where we are using social robots to promote time. inclusion between newly arrived and local chil- dren. This is part of an early career fellowship Copyright c 2018 by the author(s). awarded by the Jacobs Foundation.

7 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

We currently observe many promising and Nations, it would be (more than) great. Al- exciting advances in using AI in though there are excellent research and fund- education, going beyond automating ing initiatives in that direction, we are still not Piazza answering, how should we make there yet. use of AI to teach AI? I believe that AI can be used to complement What is your favorite AI-related movie or teachers and provide personalized instruction book and why? to students of all ages and in a variety of top- One of my favorite ones recently was the ics. Robotic tutors can play an important role Westworld TV Series because of the power in education because the mere physical pres- relationships between the human and the ence of a robot has shown to have a positive robotic characters. I find it hard to believe impact on how much information students can that humans will treat robots the way they are recall, for example when compared to a virtual treated in the series, but it makes me reflect agent displayed in a computer screen deliver- on how our future interactions with technol- ing the exact same content. ogy that is becoming more personalized and “human-like” might look like. How can we make AI more diverse? Do you have a concrete idea on what we as (PhD) students, researchers, and Help us determine who educators in AI can do to increase should be in the AI Mat- diversity our field? ters spotlight!

Something we can all do is to participate in If you have suggestions outreach initiatives targeting groups underrep- for who we should pro- resented in AI to show them that there is file next, please feel free space for them in the community. If we start to contact us via email at bottom-up, in the long-term I am positive that [email protected]. our community will be more diverse at all lev- els and the bias in opportunities, recruiting, etc. will go away.

What was your most difficult professional decision and why? Leaving my home country (Portugal) after fin- ishing my PhD to continue my research career because I miss my family and friends, and also the good weather!

How do you balance being involved in so many different aspects of the AI community? I love what I do and I currently don’t have any hobbies :-)

AI is grown up - it’s time to make use of it for good. Which real-world problem would you like to see solved by AI in the future? If AI could fully address any of the Sustainable Development Goals established by the United

8 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Events Michael Rovatsos (University of Edinburgh; [email protected]) DOI: 10.1145/3299758.3299762

This section features information about up- future Web, and enables the development and coming events relevant to the readers of AI application of intelligent technologies. WI’18 Matters, including those supported by SIGAI. features high-quality, original research papers We would love to hear from you if you are and real-world applications in all theoretical organizing an event and would be interested and technological areas that make up the field in cooperating with SIGAI, or if you have of WI. WI’18 welcomes research, application announcements relevant to SIGAI. For more as well as Industry/Demo-Track paper sub- information about conference support visit missions. Tutorial, Workshop and Special- sigai.acm.org/activities/requesting sponsor- Session proposals and papers are also wel- ship.html. come.

21st International Conference on 14th ACM/IEEE International Entreprise Information Systems Conference on Human-Robot Heraklion, Crete, Greece, May 3-5, 2019 Interaction www.iceis.org Daegu, Korea, March 11-14, 2019 The purpose of the 21st International Con- humanrobotinteraction.org/2019 ference on Enterprise Information Systems HRI 2019 is the 14th annual conference for (ICEIS) is to bring together researchers, en- basic and applied human-robot interaction re- gineers and practitioners interested in the ad- search. Researchers from across the world vances and business applications of informa- attend and submit their best work to HRI to ex- tion systems. Six simultaneous tracks will be change ideas about the latest theories, tech- held, covering different aspects of Enterprise nology, data, and videos furthering the state- Information Systems Applications, including of- the-art in human-robot interaction. Each Enterprise Database Technology, Systems In- year, the HRI Conference highlights a particu- tegration, Artificial Intelligence, Decision Sup- lar area through a theme. The theme of HRI port Systems, Information Systems Analysis 2019 is “Collaborative HRI”. The HRI confer- and Specification, Internet Computing, Elec- ence is a highly selective annual international tronic Commerce, Human Factors and Enter- conference that aims to showcase the very prise Architecture. ICEIS 2019 will be held in best interdisciplinary and multidisciplinary re- conjunction with ENASE 2019, GISTAM 2019 search in human-robot interaction with roots and FEMIB 2019. in and broad participation from communities that include but not limited to robotics, human- 2018 IEEE/WIC/ACM International computer interaction, human factors, artificial Conference on Web Intelligence intelligence, engineering, and social and be- havioral sciences. Santiago, Chile, December 3-6, 2018 webintelligence2018.com Web Intelligence (WI) aims to achieve a 11th International Joint Conference on multi-disciplinary balance between research Biomedical Engineering Systems and advances in the fields of collective intelli- Technologies gence, data science, human-centric comput- Prague, Czech Republic, February 22-24, ing, knowledge management, and network 2019 science. It is committed to addressing re- www.biostec.org search that deepens the understanding of The purpose of BIOSTEC is to bring together computational, logical, cognitive, physical as researchers and practitioners, including en- well as business and social foundations of the gineers, biologists, health professionals and Copyright c 2018 by the author(s). informatics/computer scientists, interested in

9 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 both theoretical advances and applications of information systems, artificial intelligence, sig- nal processing, electronics and other engi- neering tools in knowledge areas related to bi- ology and medicine. BIOSTEC is composed of five co-located conferences, each special- ized in a different knowledge area. BIOSTEC 2019 will be held in conjunction with MOD- ELSWARD 2019 and ICISSP 2019. Regis- tration to BIOSTEC allows free access to the MODELSWARD and ICISSP conferences (as a non-speaker).

Michael Rovatsos is the Conference Coordi- nation Officer for ACM SIGAI, and a faculty member at the Univer- sity of Edinburgh. His research is in multiagent systems and human- friendly AI. Contact him at [email protected].

10 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Conference Reports Michael Rovatsos (University of Edinburgh; [email protected]) DOI: 10.1145/3299758.3299763

This section features brief reports from recent vides a medium for presenting and exchang- events sponsored or run in cooperation with ing new scientific research ideas and techno- ACM SIGAI. logical achievements accomplished by the in- ternational community in applied intelligence. The 5th international Workshop on IEA/AIE 2018 received 146 submissions from Sensor-based Activity Recognition Canada along with 44 other countries. 64 pa- and Interaction (iWOAR 2018) pers were selected for oral presentation at the conference and included in the related pro- Berlin, Germany, September 20-21, 2018 ceedings published by Springer in the Lecture www.iwoar.org/2018 Notes in Artificial Intelligence (LNCS/LNAI) This conference-like workshop was initiated series. In addition to the papers accepted and organized by the Fraunhofer IGD and the at the main track, 22 papers were selected University of Rostock. It offers scientists, inter- (from the papers submitted directly to the spe- ested parties, and users in the area of sensor- cial session chairs) for presentation at the based activity recognition and interaction the following four special sessions and for inclu- possibility to an exchange of experiences and sion in the proceedings: Artificial Intelligence, a presentation of best-practice examples, as Law and Justice Track, Data Science, Privacy, well as technical and scientific results. The and Security Track, Intelligent Systems Ap- workshop focuses on technologies for human proaches in Information Extraction Track, and activity recognition and interaction via iner- Internet of Things and Ubiquitous Computing tial sensors (accelerometers, gyroscopes etc.) and Big Data Track. and their scientific applications. In this year’s edition of iWOAR there were 28 submissions, and the program committee chairs decided to The 15th International Conference on accept 15 papers, resulting in an overall ac- Informatics in Control, Automation and ceptance rate of 54%. For the first time at Robotics (ICINCO 2018) iWOAR, we invited 5 Posters, which were also peer-reviewed. Although the majority of au- Porto, Portugal, July 29-31, 2018 thors are based in Germany, there were also www.icinco.org/?y=2018 submissions from Austria, Denmark, Greece, ICINCO 2018 was held on Porto from 29th India, Japan, New Zealand, Spain, Sweden, to 31st July, 2018, and was sponsored by Switzerland, Turkey, and USA. the Institute for Systems and Technologies of Information, Control and Communication (IN- STICC). ICINCO 2018 was also organized The 31st International Conference on in cooperation with the ACM SIGAI, INNS, Industrial, Engineering and Other AAAI, Euromicro, Associac¸ao˜ Portuguesa de Applications of Applied Intelligent Controlo Automatico,´ euRobotics AISBL, So- Systems (IEA/AIE-2018) ciedade Portuguesa de Robotica,´ Sociedade Brasileira de Automatica´ and the Robotics So- Montreal, Canada, June 25–28, 2018 ciety of Japan. It was also technically co- ieaaie2018.encs.concordia.ca sponsored by International Federation of Au- IEA/AIE-2018 was held at Concordia Univer- tomatic Control (IFAC) and IEEE Robotics sity in Montreal, Canada this year. As in & Automation Society. ICINCO received past years, the conference was sponsored 209 paper submissions from 40 countries, by the International Society of Applied In- 22.49% of which were published and pre- telligence (ISAI), held in cooperation with sented as full papers. Four invited talks AAAI, SIGAI/ACM, and several other inter- were delivered by internationally distinguished national organizations. The conference pro- speakers, namely Keith Clark (Imperial Col- Copyright c 2018 by the author(s). lege London), Christian Schlette (Univer-

11 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 sity of Southern Denmark), and Bruno Sicil- and Technologies of Information, Control and iano (University of Naples Federico II). Ad- Communication with the Madeira Interactive ditionally, a “Best Paper Award”, a “Best Technologies Institute (MITI) as a local part- Student Paper Award” and a “Best Poster ner. The conference was also technically Award” was conferred at the conference co-sponsored by IEEE Computational Intelli- venue. More information can be found at gence Society. The conference was organized www.icinco.org/PreviousAwards.aspx. in cooperation with the ACM SIGAI, the Por- tuguese Association for Artificial Intelligence, Associazione Italiana per l’Intelligenza Artifi- The 20th International Conference on ciale, the IberoAmerican Society of Artificial Enterprise Information Systems (ICEIS Intelligence, European Society for Fuzzy Logic 2018) and Technology, the Asociacion´ Espanola˜ Funchal, Portugal, March 21-24, 2018 para la Inteligencia Artificial, the International www.iceis.org/?y=2018 Fuzzy Systems Association, European As- ICEIS 2018 (20th International Conference sociation for Artifical Intelligence, and AAAI. on Enterprise Information Systems) was held ICAART received 161 paper submissions from on Funchal, Madeira, Portugal from 21-24 34 countries, of which 27.95% were accepted March, 2018. It was sponsored by the Insti- as full papers. Four invited talks were de- tute for Systems and Technologies of Informa- livered by internationally distinguished speak- tion, Control and Communication (INSTICC). ers: Luc Steels (ICREA and UPF Barcelona), ICEIS 2018 was also organized in coopera- Virginia Dignum (Delft University of Technol- tion with the ACM SIGMIS, ACM SIGAI, ACM ogy), Eduard Hovy (CMU), and Lu´ıs Antunes, SIGCHI, AAAI, and SWIM (the IEICE Special Universidade de Lisboa, Portugal. A “Best Pa- Interest Group on Software Interprise Mod- per Award”, a “Best Student Paper Award” and elling). a “Best Poster Award” were conferred at the conference venue. More information can be ICEIS received 242 paper submissions from found at www.icaart.org/PreviousAwards.aspx 45 countries. To evaluate each submission, a double blind paper review was performed by the Program Committee. After a strin- gent selection process, 18% of the papers Michael Rovatsos is the were accepted for publication and presented Conference Coordination as full papers. Four invited talks were de- Officer for ACM SIGAI, livered by internationally distinguished speak- and a faculty member ers: Alexander Brodsky (George Mason Uni- of the School of Infor- versity), Plamen Angelov (Lancaster Univer- matics at the University sity), Salvatore Distefano (Universita´ degli of Edinburgh, UK. His Studi di Messina), and David Aveiro (Univer- research is in multiagent sity of Madeira / Madeira-ITI). systems and human- friendly AI. Contact him at The conference also awarded a “Best Paper [email protected]. Award”, a “Best Student Paper Award” and a “Best Poster Award”, which were conferred at the conference. More information can be found at: www.iceis.org/PreviousAwards.aspx

The 10th International Conference on Agents and Artificial Intelligence (ICAART 2018) Funchal, Portugal, January 16-18, 2018 www.icaart.org/?y=2018 ICAART 2018 was held on Funchal, Madeira - Portugal, from January 16 to 18, 2018. It was sponsored by the Institute for Systems

12 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

AI Education Matters: A Modular Approach to AI Ethics Education Heidi Furey (Manhattan College; [email protected]) Fred Martin (University of Massachusetts Lowell; fred [email protected]) DOI: 10.1145/3299758.3299764

Introduction Why A Modular Approach? In this column, we introduce our Model AI Here we are presenting a modular approach: Assignment, A Module on Ethical Thinking a way of incorporating an introductory lesson about Autonomous Vehicles in an AI Course, about ethics into a one-semester AI course. and more broadly introduce a conversation on Our module introduces students to the objec- ethics education in AI education. tivist framework, and opens a specific conver- sation about the ethics of self-driving cars. Why Ethics in an AI Course? The modular approach is easily integrated into a single course. If the course is popular (or re- Recognition of the need for ethics education in quired), it reaches a large portion (or all) of the engineering-related disciplines goes back the student population. Most importantly, stu- at least a hundred years (Layton Jr, 1986), but dents can connect the specific AI ideas they it’s only been since the 1990s that expecta- are learning to their ethical implications. tions for ethics education have been adopted The limitations of the modular approach are by accreditation bodies (Stephan, 1999). mostly related to its short duration—there is The use of artificial intelligence and machine only so much that can be accomplished with learning has become increasingly pervasive in only one week of instruction. our society, and this is accompanied by greatly A modular approach would be complementary expanding interest in the ethical considera- with a whole-semester course on ethical think- tions regarding their use. A crucial event was ing. Doing both would be more effective than ProPublica’s 2016 report investigating the use only one or the other. of AI tools by judges in criminal courts for de- termining the nature and lengths of sentences for convicts (Angwin, Larson, Mattu, & Kirch- ner, 2016). The article made a compelling The Model AI Assignment case that the commercial software (which is In our Module on Ethical Thinking about in widespread use and employs secret algo- Autonomous Vehicles in an AI Course, we rithms) is racially biased. shared a set of resources for faculty use: International standards bodies are also tack- ling issues related to the ethical use of AI • Two days of lecture, in-class exercises, and techniques, and specifically, the use of au- discussion, introducing Utilitarianism and tonomous vehicles. The USA published a Na- the Trolley Problem. tional Artificial Intelligence Research and De- velopment (NAIRD) Strategic Plan, and the • A requirement that the course final project EU published a report, “Recommendations paper includes a discussion of the ethical to EU Commission on Civil Law Rules on implications of the project idea. Robotics” (Martin & Makoundou, 2017). • A question on the final examination assess- Broadly, there is deep concern about the in- ing students’ understanding of the Trolley creasingly wide-reaching societal impact of AI Problem. approaches. It is urgent that our students are ready to think through the implications of their These materials are elaborated in (Furey & work, and make ethical choices. Martin, 2018); here, we add further reflec- tions and share more resources for introduc- Copyright c 2018 by the author(s). ing ethics to AI students.

13 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Reflections on Ethics Teaching It is useful to give them some general back- ground in the field of ethics. For instance, It is One challenge to incorporating ethics into an important for students to recognize that there AI course is helping students recognize the are different sorts of questions that one might structure of ethical problems and their solu- ask regarding human behavior, only some of tions. The field of ethics is both complex and fall under the domain of ethical inquiry. Ethi- far-ranging. Because of this, it can be difficult cists, as opposed to psychologists or soci- to meet this challenge within the space of an ologist, are interested in what people ought AI course without getting too far a field of the morally do to apart from what they in fact do standard course material. With this concern or why they do it. in mind, we selected a topic that could serve as an example both of an ethical problem and ethical problem solving: the ethics of algorithm Other Resources development for autonomous vehicles. Burton et al.’s “Ethical Considerations in Ar- In the module, students were first introduced tificial Intelligence Courses” (2017) provides to the connection between ethical algorithms an in-depth introduction to this conversation, for autonomous vehicles and a classic eth- including case studies and resource links. ical dilemma: the Trolley Problem. After- These authors also contribute the idea of us- word, students were guided through a work- ing science fiction to engage students in think- sheet driven group exercise designed to fos- ing through ethical considerations (Burton, ter discussion and debate on the topic. This Goldsmith, & Mattei, 2015). discussion helped prepare students to eval- uate the potential benefits and shortcomings For a short video introduction to some ethics of an intuitive solution to ethical problems— issues in AI, see Atlantic Magazine’s Moral Utilitarianism. Finally, students were offered Code: The Ethics of AI. examples of how one might construct a solu- For a variety of AI ethics resources including tion to these problems and challenged to con- research, symposia, workshops, and reports, tinue thinking about the issue. visit the AI Now Institute at New York Univer- sity. Thoughts on Implementation The AI Ethics Lab is a virtual organization that “brings together researchers and practitioners The Trolley Problem presents a seeming in- from various disciplines to detect and solve is- tractable ethical dilemma — one in which sues related to ethical design in AI.” every solution comes with an ethical cost. Several articles from popular media address One benefit of exposing students to a some- the challenges of introducing AI ethics into the what eccentric philosophical example is that classroom. For instance, see (Tugend, 2018) students become familiar with a key ethical and (Holmes, 2018). problem-solving tool — the use of “thought experiment” — highly idealized hypothetical To date, more research has been done in en- cases designed to test ethical theories and gineering ethics pedagogy than in AI ethics to isolate relevant moral variables. Another pedagogy. For this reason, it can be use- benefit is that students come to understand ful to look for resources in engineering ethics. the complex nature of ethical dilemmas, and For a problem-solving approach to ethics in- to recognize that solutions to such problems struction, see (Whitbeck, 1996). For a brief are rarely straightforward and may perhaps overview of some of the major ethical theories be equally complex. Too often, students who along with a discussion benefits and difficul- are unfamiliar with ethical problem solving re- ties of teaching ethical theory, see (Bouville, sist thinking about ethical dilemmas because 2008). For an article on teaching higher-order the answers appear “unknowable.” Here, it is ethical concepts in engineering, see (Haws, useful to draw parallels between difficult moral 2004). For an example of an alternative to questions and difficult technical questions, re- stand-alone ethics instruction in engineering, minding them that that complexity does not see (Riley, Davis, Jackson, & Maciukenas, necessarily equal intractability. 2009).

14 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

References Heidi Furey is an As- sistant Professor of Angwin, J., Larson, J., Mattu, S., & Kirchner, L. Philosophy and Director (2016, May). Machine bias: There’s soft- of the Ethics Center at ware used across the country to predict at Manhattan College. future criminals. And it’s biased against Furey researches applied blacks. ProPublica, 23. ethics and ethical peda- Bouville, M. (2008). On using ethical theories gogy. Her current projects to teach engineering ethics. Science and include work on the ethics Engineering Ethics, 14(1), 111–120. of autonomous vehicles. Burton, E., Goldsmith, J., Koenig, S., Fred Martin is Profes- Kuipers, B., Mattei, N., & Walsh, T. sor of Computer Science (2017). Ethical considerations in artifi- and Associate Dean for cial intelligence courses. arXiv preprint Student Success in the arXiv:1701.07769. Kennedy College of Sci- Burton, E., Goldsmith, J., & Mattei, N. (2015). ences at the University Teaching AI ethics using science fiction. of Massachusetts Low- In Aaai workshop: Ai and ethics. ell. Martin develops novel Furey, H., & Martin, F. (2018). Introducing eth- technologies for CS and ical thinking about autonomous vehicles data science education, particularly for K–12. into an AI course. In AAAI-2018. Haws, D. R. (2004). The importance of meta- ethics in engineering education. Science and Engineering Ethics, 10(2), 204– 210. Holmes, W. (2018). The ethics of artificial in- telligence in education. University Busi- ness. Layton Jr, E. T. (1986). The revolt of the engi- neers. social responsibility and the amer- ican engineering profession. ERIC. Martin, C. D., & Makoundou, T. T. (2017). Tak- ing the high road ethics by design in AI. ACM Inroads, 8(4), 35–37. Riley, K., Davis, M., Jackson, A. C., & Maciukenas, J. (2009, March). ”ethics in the details”: Communicating engineering ethics via micro-insertion. IEEE Trans- actions on Professional Communication, 52(1), 95-108. doi: 10.1109/TPC.2008 .2012286 Stephan, K. D. (1999). A survey of ethics- related instruction in U.S. engineering programs. Journal of Engineering Edu- cation, 88(4), 459-464. Tugend, A. (2018). Colleges grapple with teaching the technology and ethics of A.I. New York Times. Whitbeck, C. (1996). Ethics as design: Do- ing justice to moral problems. Hastings Center Report, 26(3), 9–16.

15 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

AI Policy Matters Larry Medsker (George Washington University; [email protected]) DOI: 10.1145/3299758.3299765

Abstract Board, a bipartisan agency within the execu- tive branch. He will serve as a part-time mem- AI Policy is a regular column in AI Matters ber of the board while continuing his teaching featuring summaries and commentary based and research at Princeton. The five-person on postings that appear twice a month in the board is charged with evaluating and advising AI Matters blog (https://sigai.acm.org/aimatters/ on executive branch anti-terrorism measures blog/). Selected posts are summarized in is- with respect to privacy and civil liberties. It is sues of AI Matters. a very important issue, Felten said. Federal agencies, in the course of doing national se- Introduction curity work, have access to a lot of data about The SIGAI Public Policy goals are to people and they do intercept data. Its impor- • promote discussion of policies related to AI tant to make sure they are doing those things through posts in the AI Matters blog on the in the way they should and not overstepping. 1st and 15th of each month, Felten added that the board has the authority • help identify external groups with common to review programs that require secrecy. The interests in AI Public Policy, public has limited visibility into some of these • encourage SIGAI members to partner in programs, Felten said. The board’s job is to policy initiatives with these organizations, look out for the public interest. and • disseminate public policy ideas to the SIGAI membership through articles in the newslet- FTC Hearings on AI ter. On November 13 and 14, the Federal Trade I welcome everyone to make blog comments Commission (FTC) held the seventh hearing so we can develop a rich knowledge base of in its series of nine planned Hearings on Com- information and ideas representing the SIGAI petition and Consumer Protection in the 21st members. Century. From the FTC: ”The hearing will ex- amine competition and consumer protection issues associated with the use of algorithms, AI Appointments artificial intelligence, and predictive analytics The Computing Community Consortium in business decisions and conduct. See de- (CCC) announced a new initiative to create tailed agenda. The record of that proceed- a Roadmap for Artificial Intelligence. SIGAI’s ing will be open until mid-February. To further Yolanda Gil (University of Southern California its consideration of these issues, the agency and President-Elect of AAAI) will work with seeks public comment on the questions, and Bart Selman (Cornell University) to lead the it welcomes input on other related topics not effort. The initiative will support the U.S. Ad- specifically listed in the 25 questions.” See the ministration’s efforts in this area and involve agenda and more information in the press re- academic and industrial researchers to help lease. Comments can be submitted online no map a course for needed research in AI. They later than February 15, 2019. will hold a series of workshops in 2018 and 2019 to produce the Roadmap by Spring of AI Safety By Design 2019. Policymakers need to think of AI and Au- Ed Felten, Princeton Professor of Computer tonomous Systems (AI/AS) as always need- Science and Public Affairs, has been con- ing varying degrees of the human role (hy- firmed by the U.S. Senate to be a member of brid human/machine systems). Understand- the U.S. Privacy and Civil Liberties Oversight ing the potential and limitations of combining Copyright c 2018 by the author(s). technologies and humans is important for re-

16 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 alistic policymaking. A key element, along with perience in corporate law and contract review accurate forecasts of the changes in technol- were pitted against an AI. ”Their task was ogy, is the safety of AI/AS-Human products as to spot issues in five Non-Disclosure Agree- discussed in the IEEE report Ethically Aligned ments (NDAs), which are a contractual ba- Design, which is subtitled “A Vision for Prior- sis for most business deals.” The LawGeex itizing Human Wellbeing with Artificial Intelli- AI system attempted correct identification of gence and Autonomous Systems”, and Ben basic legal principles in contracts The re- Shneidermans excellent summary and com- sults suggest that AI systems can produce ments on the report as well as the YouTube higher accuracy in shorter times compared video of his Turing Institute Lecture on “Algo- to lawyers. As with other areas of AI appli- rithmic Accountability: Design for Safety”. cations, issues include trust in automation to make skilled legal decisions, safety in using In Shneidermans proposal for a National Al- AI systems, and impacts on the workforce of gorithms Safety Board, he writes What might the future. For legal work, AI systems poten- help are traditional forms of independent over- tially reduce the time needed for high-volume sight that use knowledgeable people who and low-risk contracts and give lawyers more have powerful tools to anticipate, monitor, and time to work on less mundane work. Policies retrospectively review operations of vital na- should focus on automation where possible tional services. The three forms of inde- and safe, and the AI for legal work is another pendent oversight that have been used in example of the need for collaborative roles for the past by industry and governmentsplanning human and AI systems. oversight, continuous monitoring by knowl- edgeable review boards using advanced soft- ware, and a retrospective analysis of disaster- AI Impact on Litigation: Litigation is emerging sprovide guidance for responsible technology in all parts of society related to the use of AI. leaders and concerned policy makers. Con- Understanding the nature of adaptive AI sys- sidering all three forms of oversight could lead tems can be crucial for fact-finders and diffi- to policies that prevent inadequate designs, bi- cult to explain to non-experts. Smart policy- ased outcomes, or criminal actions. making needs to make clear the liability is- sues and ethics in cases involving the use Efforts to provide safety by design include of AI technology. Artificial Intelligence and work at Google on Human-Centered Machine the Role of Expert Witnesses in AI Litigation Learning and a general human-centered ap- by Dani Alexis Ryskamp, writing for The Ex- proach “that foregrounds responsible AI prac- pert Institute, discusses artificial intelligence tices and products that work well for all peo- in civil claims and the role of expert witnesses ple and contexts. These values of responsible in elucidating the complexities of the technol- and inclusive AI are at the core of the AutoML ogy in the context of litigation. “Over the past suite of machine learning products ”. Fur- few decades, everything from motor vehicles ther work is needed to systemize and enforce to household appliances has become more good practices in human-centered AI design complex and, in many cases, artificial intelli- and development, including algorithmic trans- gence only adds to that complexity. For end- parency and guidance for selection of unbi- users of AI products, determining what went ased data used in machine learning systems. wrong and whose negligence was responsible can be bafflingly complex. Experts retained in AI cases typically come from fields like Legal AI computer or mechanical engineering, informa- AI is impacting law and policy issues as both tion systems, data analysis, robotics, and pro- a tool and a subject area. Advances in AI pro- gramming. They may specialize in questions vide tools for carrying out legal work in busi- surrounding hardware, software, 3D-printing, ness and government, and the use of AI in all biomechanics, Bayesian logic, e-commerce, parts of society is creating new demands and or other disciplines. The European Com- challenges for the legal profession. mission recently considered the question of whether to give legal status to certain robots. Lawyers and AI Tools: In a recent study, “20 One of the issues weighed in the decision in- top US corporate lawyers with decades of ex- volved legal liability: if an AI-based robot or

17 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 system, acting autonomously, injures a per- sirable that you fill it out for all authors. This son, who is liable?” information will be deleted from EasyChair af- ter the conference.” Pew Research Report We recommend that to evaluate the likelihood of privacy protection, one should pay atten- Pew Research Center recently released a re- tion to the EasyChair Terms of Service, par- port “Public Attitudes Toward Computer Algo- ticularly Section 6 “Use of Personal Informa- rithms”, by Aaron Smith, on Americans con- tion” to assess better the level of risk if our cerns about fairness and effectiveness in mak- members choose to enter personal informa- ing important decisions. The report says “This tion. Your Public Policy Officer recommends skepticism spans several dimensions. At a further discussion among the SIGAI member- broad level, 58 percent of Americans feel that ship about requesting changes in ACM policy. computer programs will always reflect some Please send your views on this topic to SIGAI level of human bias although 40 percent think officers and contribute comments to the policy these programs can be designed in a way that blog. is bias-free. And in various contexts, the pub- lic worries that these tools might violate pri- vacy, fail to capture the nuance of complex situations, or simply put the people they are Larry Medsker is Re- evaluating in an unfair situation. Public per- search Professor of ceptions of algorithmic decision-making are Physics and was founding also often highly contextual the survey pre- director of the Data Sci- sented respondents with four different scenar- ence graduate program at ios in which computers make decisions by col- The George Washington lecting and analyzing large quantities of pub- University. Dr. Medsker lic and private data. Each of these scenarios is a former Dean of the were based on real-world examples of algo- Siena College School of rithmic decision-making and included: a per- Science, and Professor sonal finance score used to offer consumers in Computer Science and in Physics. His deals or discounts; a criminal risk assess- research in AI includes work on artificial ment of people up for parole; an automated neural networks, hybrid intelligent systems, a resume screening program for job applicants; and the impacts of AI on society and policy. . and a computer-based analysis of job inter- He is the Public Policy Officer for the ACM views. The survey also included questions SIGAI. about the content that users are exposed to ahttp://humac-web.org/ on social media platforms as a way to gauge opinions of more consumer-facing algorithms.”

EasyChair Policy Issue We are revisiting an issue raised last June in- volving data privacy for SIGAI and ACM mem- bers using EasyChair to submit articles for publication, including the AI Matters Newslet- ter. As part of entering a new submission through EasyChair, the following message ap- pears: “AI Matters, 2014-present, is an ACM conference. The age and gender fields are added by ACM. By providing the information requested, you will help ACM to better under- stand where it stands in terms of diversity to be able to focus on areas of improvement. It is mandatory for the submitting author (but you can select prefer not to submit) and it is de-

18 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Epochs of an AI Cosmology Cameron Hughes (Northeast Ohio ACM Chair; [email protected]) Tracey Hughes (Northeast Ohio ACM Secretary; [email protected]) DOI: 10.1145/3299758.3299868

Epochs and Hype Cycles and Mathematical Logic Machines. We could ask ourselves what causes one AI epoch to We began the discussion about the educa- start and another to end? We might look at an tional benefits to the general public of a stan- AI epoch from the points of view of (research, dard AI Cosmology in Volume 4 Issue 2 of “AI investment, utilization) funding, media, indus- Matters”. But the mere notion of such a stan- try, and social impact. For critical or perhaps dard immediately raises lots of questions. In even cynical purposes, we might ask how or this column, we only focus on a very narrow if AI epochs correspond or correlate to tech- set of those questions: First and foremost: nology hype cycles. Figure 1 shows how a Is there an AI Cosmology? If in fact the no- generic tech hype cycle curve. tion of cosmology applies to AI, is there only one valid AI Cosmology or are there compet- ing valid cosmologies? If there is only one valid cosmology, what is its structure? How far can we take an AI Cosmology analogy? Like our Universe, did AI have the equivalent of a Big Bang? Is the field or structure of AI ex- panding or contracting? In an interview with Nikola Danaylov, podcaster of Singularity.FM, the late Marvin Minsky1, one of the early AI pioneers, suggested that the evolution of AI in the 21st century maybe be contracting or even stalled! We posit that an AI Cosmology would help to educate the general public about our research and work, and its impact on society. Internally, it would help us to fix the objects, structure, and dynamics of the field. A stan- dard cosmological model would help us all dis- tinguish spurious AI mythos, pseudo science, hype, and fake news from the genuine arti- Figure 1: A Generic Tech Hype Curve. cle. A true cosmology of Artificial Intelligence would help us fix its origins, understand its tra- Notice that a hype cycle reaches a peak where jectory, and predict its fate. expectations are unrealistically high, the pub- We are currently at the beginning of what we lic is at its height of misinformation, and claims could consider the epoch of Machine Learn- of technologys applications are inflated. Fol- ing. Of course how the field will ultimately lowing the curve in Figure 1 is the realiza- measure the epochs of AI remains an open tion that the inflated claims are not coming question. But dividing the history of AI into true, funding and investment decreases, and epochs might help us better understand the then the public becomes disillusioned. Is it cosmology of AI a little better. For instance, the case that AI epochs happen to coincide we could say that before the epoch of Machine with AI hype cycles? Are AI hype cycles typi- Learning was the epoch of Expert Systems, cally components of epochs? Are they one in and prior to that, the epochs of Cybernetics the same? One of our current interests with respect to the general public’s vulnerability to Copyright c 2018 by the author(s). AI hype cycles, how much do technical termi- 1https://www.youtube.com/watch?v= nology misnomers, misconstrued AI concepts, 3PdxQbOvAlI and ill-defined AI vocabulary contribute? In

19 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Figure 2, we show our first rough draft of AIs • Scene analysis Epochs we’ve discussed so far. • Natural vs. Artificial Intelligence • Theorem Proving • Implementing A Question-Answering sys- tems

AI Magazine Volume 3 Number 1 (1982) re- ports that AI and Brain Theory at UMASS was still intermingling the notions of AI and Cyber- netics. Our program in AI is part of the larger depart- Figure 2: Rough Draft of AI’s Epochs. mental focal area of Cybernetics which inte- grates both AI and Brain Theory (BT). Our re- search also draws upon a new and expanding In Figure 2, each AI epoch has a timeline and interdepartmental program in Cognitive Sci- within each epoch, there has been a hype cy- ence that brings together researchers in Cy- cle. For now we conjecture that hype cycles bernetics, Linguistics, Philosophy, and Psy- and epochs are not the same. Epochs start chology. prior to the hype cycle and continues after the hype cycle is over. The same fervor that Pop Culture now has for Machine learning was once had for Expert Systems. Prior to Expert Systems, feedback Machine Learning and Back to communication and control in animals and Cybernetics machines was all the craze and we were in the epoch of Cybernetics. It was just like the par- We are in the Machine Learning epoch. In AI allels between the AI/Machine Learning and Matters Volume 4 Issue 2, we talked about the AI/Expert System explosions. Cybernet- the age of Expert Systems. But before AI ics was the talk of the town complete with sci- manifested itself in the form of Expert Sys- entific breakthroughs, industrial innovations, tems, Cybernetics represented the promise of massive social change and a strong dose Artificial Intelligence. In June 1975, our own of Cyber-Techno-Hype. In 1969, Georges SIGART Newsletter No. 52 contained an arti- Boulanger, the president of the International cle by Edward M. Riseman entitled “AAI AND Association of Cybernetics noted that “To the BRAIN THEORY AT THE UNIV. OF MAS- general public cybernetics conjures up visions SACHUSETTS”. Riseman defines Cybernet- of some fantastic world of the future peopled ics accordingly: by robots and electronic brains” (Boulanger, At the University of Massachusetts at 1969). Cybernetics represented the embod- Amherst, we use the name to describe iment, incarnation, goals, aspirations, and the study of Artificial Intelligence (AI) and funding targets of some of the same subject Brain Theory (BT). At the University of matter that is reserved for AI/Machine Learn- Massachusetts in 1975, the Computer Sci- ing. Cybernetics was a staple of science fic- ence and Information Science Program was tion and a fad among artists, musicians, and roughly divided into three major areas: Com- intellectuals in the 1950s and 1960s. These puter Systems, Theory of Computation and facts are well documented in Ronald Kline’s Cybernetics. 2015 tour de force book “The Cybernetics Mo- ment” (Kline, 2015) where he painstakingly In Riseman’s SIGART article, he describes captures the history of Cybernetics, Informa- the 1974-1975 graduate courses in the Fun- tion Theory, and the early AI Ethos. Upon a damentals of Cybernetics as including: close reading of Kline’s “The Cybernetics Mo- ment”, one can almost see how the Artificial • Computational Cybernetics Intelligence baton would eventually pass from • Artificial Intelligence Cybernetics to Expert Systems. • Neurocybernetics of the Cerebral Cortex The intersection between the actual artifacts

20 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 of AI research and the general publics un- ute to Dr. Norbert Weiner, ed. J. Rose. Lon- derstanding and expectations with regard to don, England. Iliffe Books. those artifacts can be perilous. It appears that this has been the case for each of the AI epochs! As AI researchers, educators, and practitioners, we have an obligation to Cameron Hughes is a minimize (as best we can) the misconstruc- computer and robot pro- tions, misunderstandings, and misrepresenta- grammer. He is a Soft- tions of our work and our field. At Ctest Lab- ware Epistemologist at oratories, we are positing the notion of de- Ctest Laboratories where vising a “Standard Model” for an AI Cosmol- he is currently working on ogy that would standardize the structure, tax- A.I.M. (Alternative Intelli- onomy, and ontology in a form that is suit- gence for Machines) and able for consumption by early college STEM A.I.R (Alternative Intelli- programs, high schools, and the general pub- gence for Robots) tech- lic. The moniker AI remains for much of the nologies. Cameron is general public a reference to vague, ambigu- the lead AI Engineer for ous, futuristic, even threatening subject mat- the Knowledge Group at ter. This was also the case for the moniker Advanced Software Con- Cybernetics. But after all what is cybernetics? struction Inc. He is a member of the advisory Or rather what is it not, for paradoxically the board for the NREF (National Robotics Edu- more people talk about cybernetics the less cation Foundation) and the Oak Hill Robotics they seem to agree on a definition, George Makerspace. He is the project leader of the Boulanger 1969 (Boulanger, 1969). technical team for the NEOACM CSI/CLUE In this column, we would like to instigate a Robotics Challenge and regularly organizes conversation with our fellow SIGAI members and directs robot programming workshops for on this notion of a standard model for AIs Cos- varying robot platforms. Cameron Hughes mology. So far we’ve offered conjectured no- is the co-author of many books and blogs tions of standardized epochs of AI, AI Hype on software development and Artificial Intelli- Cycles, standard AI taxonomies, and a stan- gence. dard AI domain ontology as a starting point. Tracey Hughes is a soft- What do we really mean by “Artificial Intel- ware and epistemic visu- ligence”? If the evolution of Artificial Intelli- alization engineer at Ctest gence can be divided into epochs, where are Laboratories. She is we now, and what was the first one? Can the lead designer for the we understand the evolution of Artificial Intelli- MIND, TAMI, and NO- gence, its technologies and applications in the FAQS projects that uti- context of a AI cosmological standard model? lize epistemic visualiza- If we can construct an AI cosmological stan- tion. Tracey is also a dard model, can we represent it in way that is member of the advisory suitable for consumption by the general pub- board for the NREF (Na- lic? If the epoch of Cybernetics preceded the tional Robotics Education epoch of Expert Systems, was there an AI Foundation) and the Oak epoch that preceded Cybernetics? We’ll try Hill Robotics Makerspace. to fill in more of the timeline in the next article. She is the lead researcher of the technical team for the NEOACM CSI/CLUE Robotics References Challenge. Tracey Hughes is the co-author with Cameron Hughes of many books on soft- Kline, R. (2015). The Cybernetics Moment. ware development and Artificial Intelligence. Baltimore, MD, USA: John Hopkins University Press. Boulanger, G.R. (1969) Prologue: What is Cy- bernetics?, in Survey of Cybernetics: A Trib-

21 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

Artificial Intelligence and Jobs of the Future: Adaptability Is Key for Human Evolution Sudarshan Sreeram (Vidya Mandir Senior Secondary School, Mylapore, Chennai - 04, India; [email protected]) DOI: 10.1145/3299758.3300060

Humans, the most intelligent species on events is influenced by two factors our mind- planet Earth, have always found a way to set and our reaction towards change. For adapt to their dynamic environment and sur- example, consider an article published by vive in a consistent manner. We have made Columbia University on the psychology of tremendous progress in evolving from primi- gullible mindsets (Valhouli, n.d.). According to tive men and women in small, dreary caves this publication, a fourteen-year-old boy con- to modern men and women in large, bustling vinced forty-three of fifty people to support a cities. Our passionate strive to survive and ban on water. The boy referred to water as affinity to adapt is similar to that shown by dihydrogen monoxide, the chemical name of mosquitoes; regardless of the continuous de- water, in a petition that was circulated through velopment and massive deployment of in- the internet. An important factor to consider sect repellent products, these undeniably irk- is the way this petition was framed to influ- ing insects, which cause a plethora of life- ence human minds; he claimed that dihydro- threatening diseases, always make a striking gen monoxide was a harmful chemical that comeback (Stanczyk, Brookfield, Field, & Lo- was found in tumors of cancer patients. Sim- gan, 2013). ilar convincing sentences use widely known terms such as sweating, vomiting, erosion, Humans have a natural tendency to be preoc- and lethal to lure an uninformed reader. The cupied with what interests us the most. We reader takes the authenticity of the provided lack the ability to not think of anything, even information for granted without checking facts for a second. Boredom is not an option for hu- or running background checks on its source. mans; this is evidenced by a research exper- The information is then dispersed through ca- iment conducted in 2014 by Timothy Wilson, sual conversations and discussions to a wider a social psychologist professor at the Univer- network, leading to the general public be- sity of Virginia, Charlottesville (Wilson et al., ing falsely informed. Therefore, it is evident 2014). The results of this research found that that our mindsets are one of the most gullible humans would go to the extent of shocking and volatile entities on our planet; widespread themselves with small jolts of electricity, when controversies, futuristic sci-fi movies, factu- locked alone in a lab, just to keep themselves ally inaccurate data, and popularized opinions occupied. Our minds were built to be aware of highly influential representatives do a very of the world around us and engage in actions good job at molding it. and decisions that shape it. Thus, it is not pos- sible to stop continuously flowing thought pro- Change-events are stimuli for human soci- cesses (Whitehead, 2014). eties. When the change is for the better and provides visible benefits, the reaction is most Our rich and vivid history may be described likely positive. For example, the establishment as a continuously flowing river, which does not of independence or democracy in a country, heed to a specific, predictable pattern. How- the creation of new job opportunities, etc. In ever, the river does take similar looking turns a similar way, when the change is disruptive as it flows. Applying this hypothetical case to and removes certain key defining aspects of real world history, we see that major events, human life, the reaction is most likely negative. be it global revolutions, wars, famines, natural For example, the increase in income taxes and disasters, and even plagues, do recur, though, fuel prices or the decrease in job satisfaction and at irregular intervals. leading to lack of peaceful life (Rohani & Pa- The time interval between two successive hazri, 2018). Copyright c 2018 by the author(s). Industrial revolutions, dubbed as the perma-

22 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 nent markers of change, have had monumen- The reason behind these misconceptions and tal impacts, both positive and negative, on the controversies is due to a misunderstanding of lives of humans. An industrial revolution can one word Technology. Technology is a life- be thought of as a steering wheel of a car that altering word that has found its place in hu- represents the human civilization; this car has man society. Over the past few decades, it made three sharp turns and is on a straight has been integrating itself into all aspects of road, looking forward to making a fourth turn. human life, and now, it is integral to our daily lifestyles. Our connection with technology has The first turn, which had the largest radius, grown so deep that we cannot even fathom an witnessed the urbanization and industrializa- alternate timeline in which humans live without tion of several rural societies as well as the it. Although widely debated, its intended pur- deployment of steam engines, which played a pose is to make human life simple and easy. It major role in the development of transporta- helps us achieve milestones and immensely tion technologies during the 18th and 19th speeds up the progress of our society as a centuries. The second turn saw the growth whole. of industrial centers around the world right before the first world war and the introduc- We have witnessed the rapid transformation tion of mass-produced electrical appliances, of our landscapes and interests as technology which included technological advancements molds the shape of our future. It functions as that would change the world forever. The third a tool of change that teaches us how to adapt turn, which started during the early 1980s, and evolve to meet the requirements of future took the car onto the digital highway; the dig- societies. Today, some of us question these ital revolution and the boom of the internet requirements and are doubtful of what this fu- made it possible for almost everyone in the ture might hold. It is easier and faster for hu- world to stay connected. The fourth turn, the mans to find flaws in an idea or a concept than most anticipated of them all, deals with signifi- it is to understand its true value. cant advancements in the fields of 3D printing (Kyle, 2018), bio-circuits (Prox, Smith, Holl, Humans have shown varied reactions to the Chehade, & Guo, 2018), quantum comput- use of automation technology and robotics ing (Chong, 2018), robotics, IoT (Agiwal, Sax- in several industrial areas (Knight, 2018). ena, & Roy, 2018), nanotechnology (Alvarez, Consider the widely discussed transition of Chan, Elimelech, Halas, & Villagran´ , 2018), the manufacturing industry from recruiting renewable energy technologies, self-driving skilled human workers to employing auto- vehicle technologies (Bansal & Kockelman, mated robotic technologies. The first flaw hu- 2018), and finally, the most controversial and mans see with this transition is the massive popular of them all Artificial Intelligence (AI) unemployment of skilled laborers. What we (Lu, Li, Chen, Kim, & Serikawa, 2018). fail to see is that this transition not only pre- vents us from working in hazardous conditions Ever since its inception in 1956, when John but also reduces a companys expenditure in McCarthy held a conference on the subject manufacturing processes. In addition, the end (McCarthy, Minsky, Rochester, & Shannon, product is of a much higher quality due to re- 1955), Artificial Intelligence has always caught duction in human errors, and the time taken the imagination of those who take it upon by the different processes is significantly re- themselves to predict the outline of future so- duced, thereby increasing production rates to cieties (Barrat, 2015). It is undoubtedly an meet the demand. Yet, despite all these ad- inspiring and captivating field; however, it is vantages, we still consider the loss of jobs as met with a great deal of controversies and a prominent result. misconceptions regarding its purpose and in- tention. Some research suggests controver- Humans have always developed a notion of sial ways in which it would influence human resistance and anxiety over advanced devices life while others see the potential for a future and technologies, be it self-driving cars or arti- where AI systems complement human soci- ficial intelligence. Even in the 1980s, the term eties. An oversimplified and partially misun- computerphobia was widely circulated in mag- derstood influence of AI is its effective take azines, newspapers, and psychology studies over of human-centralized jobs in the future. (Lafrance, 2015). The term was mainly used

23 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 to address a persons fear of being threat- ing at a snails pace. Marlene Behrmann, a ened and replaced by a computer. In addi- cognitive neuroscience professor at Carnegie tion, people felt hostile in the presence of a Mellon University, claimed that neuroscientists computer. Eventually the presence of comput- only have a primitive understanding of human ers in our daily lives grew exponentially. This brain functions despite decades of research is quite synonymous to what is going on in (Gornall, 2014). Thus, when humans them- the present day with artificial intelligence tech- selves have not yet understood the complete nologies. The general trend marks a peak working of their brains, it is improbable to cre- period of hype and gradually settles into the ate an intelligent system that can replicate it. books of history. AI systems are excellent at performing repet- The Gartner hype cycle (Hype Cycle Re- itive tasks, which is evidenced by their search Methodology, n.d.) for emerging tech- purpose-built nature. It can automate almost nologies in 2018 depicts the presence of arti- any system that is repetitive or predictable. ficial intelligence on two of the five regions in Thus, it is not surprising to see the results of the hype cycle graph innovation trigger and AI technologies from Google and IBM (Hiner, peak of inflated expectations. In addition, the 2018; Russell, 2017; Best, n.d.). Even if AI graph also predicts that these various tech- were to advance to match the human intellect, nologies under AI may reach the plateau of their only way of physically interacting with this productivity a period of mainstream adapta- world is through robotic bodies. Even the high- tion in roughly ten or more years (Panetta, est quality of robots are prone to failures, given 2018). that their developments should be cost effec- Ever since its creation, a computer was tasked tive. Regardless, robotic technicians will be with replicating and efficiently performing hu- required to diagnose and address any techni- man tasks. It is just human notion that when a cal issues. Thus, there will always be space computer is able to do something new, we re- for humans. gain those feelings of anxiety and resistance. Moreover, educational institutions around the People, who are not properly informed of a world are offering courses on robotic systems specific technology, are easily persuaded by to advance the technical skills required to work the depictions of AI as a threat. So, what is in this field. In fact, there is pressure on Artificial Intelligence? the educational sector to teach students to Todays AI is a system of complex algorithms code from a young age (Barone, 2017). Sev- that do an exceptional job at processing and eral technology giants, including Apple, are analyzing large volumes of data; it is a skill bringing forward their support to educate chil- that humans lack due to the volatility and com- dren on coding practices (Education - Teach- putational limitations of their memory systems. ing Code, n.d.). Since the world is moving AI systems are purpose built for executing forward with technology as its backbone, the specific tasks, so there are several severe lim- people living in it are expected to be well in- itations on what it can accomplish; to put it in formed of it in a transparent manner. simple terms, AI is not yet anywhere near the ability to freely think as we do. As for the role of Artificial Intelligence (AI) in jobs of the future, AI systems would en- Super intelligence or true AI, on the other gage in cooperation with humans; these intel- hand, is a form that current AI technolo- ligent systems would provide humans with rel- gies could possibly take in the future as we evant data and suggestions to solve a particu- gain a deeper knowledge regarding its fun- lar task. There is a clear line that distinguishes damental concept the working of the hu- jobs and tasks when considering the potential man brain (Hassabis, Kumaran, Summerfield, of current AI technologies (Pring, 2018). A job & Botvinick, 2017). This future is not com- is a simple term for the human state of being ing anytime soon; currently, there is no one preoccupied with multiple tasks while benefit- on our planet who actually understands how ing from a source of income; whereas, a task super intelligence works (Byttner, 2017). Its is a piece of work that acts as a subset of a progress is most entirely dependent on human job. It is a definite possibility that in the future, brain research, which is currently progress- AI systems would complement the human job

24 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 workload rather than completely making it ob- is not true in the real world as its dynamics solete. hold innumerable factors. Consider the case where a robotic machine breaks down in a It is evident and inevitable, though, that new large scale factory. A human technician is re- fields of interest to humans will rise as the quired at this point for analyzing the situation depths of our knowledge in existing fields in- and factoring multiple variables and carrying crease. Our future is influenced by a great out an in-depth root cause analysis. Humans deal of factors, AI being just one of them. Con- are better at decision making as they have a trary to the notion that AI would make us un- lot more input and background knowledge to employed in the future, this intelligence would, work with and consider, which are not just de- in actuality, boost job growth; though, the vari- rived from text book sources (Why humans ety of jobs offered in the future may not be the learn faster than AI-for now., 2018). Critical same as what is offered in the present day. situations such as fire emergencies and natu- For example, humans may expect titles such ral disaster management, where even robots as organ fabricator, commercial space pilot, arent effective, do require human presence alternate energy consultant, and trash engi- (Why humans learn faster than AI-for now., neer (9 Top Jobs in 2030: Future Skills You 2018). Unless super intelligence is reached, Need To Learn Now, n.d.). Prior to advance- AI is nothing all by itself apart from being a ments in AI, it would have been hard to imag- follower of specific instructions given by hu- ine job titles such as Machine Learning Devel- mans. To reach a future where AI would sup- oper or Artificial Intelligence Specialist. Using port job growth and cooperate with humans, the same analogy, we may expect several new three fundamental principles must be estab- job titles in the future as technology evolves. lished: transparency in actions and decisions, restriction on information collection and ac- As jobs evolve over time so do the tools we cess of data sources, and foundation for long- use. This is evidenced by the history of art; term reliability and trust. during the 1800s, people had the time and pa- tience to sit for days, months, or even years In the far future, we may reach a state when to complete a single painting, made entirely artificial intelligence will prioritize and reorga- by hand. Today, we take aid of digital drawing nize our skillsets; humans will think but not do. programs and highly precise graphics pen dis- The creation of objects and physical comple- plays or tablets to speed up that process and tion of tasks will be taken care by humanoids express our thoughts as they flow. Though, powered by intelligent AI, as shown in film this example does bring into question the adaptations of popular sci-fi novels. Contrary value of art and the presence of human touch, to what is shown in these films, humanoids which, in my opinion, would fade as we move are less likely to turn against humanity as the into the future. However, we do not surren- three principles, coupled with safety protocols der our creativity and intellectual abilities, but and other measures, would be enforced upon express the same in a different form that con- them. Humans would also have to make sure sumes less time and effort. Artists of the 19th that the AI or AI based systems cannot find century had brushes and canvases, artists of a way to bypass these security measures and the 21st century have graphics tablets and gain access to unauthorized sections of data digital canvases, and artists of the future may and services. In the same way, the general have technologies that combine the use of AI public should also be restricted from gaining and Augmented Reality (AR), supported by privileges to access the core of the AI system. the human brain, to provide everyone with an This would keep intelligence away from peo- artistic toolset. This shows that creativity is ple who would utilize it for social misconduct, a human characteristic that cannot simply dis- disinformation, and to promote corruption. appear in the presence of artificial intelligence. With these security measures in mind, hu- A research paper from Oxford university (Frey mans can look forward to a future filled with & Osborne, 2017), which claims that AI would cooperation and responsibility. In addition, take over human jobs of the future, bases its with heavy investments in nuclear fusion and results on assuming that all jobs are repeti- space mining research (James, 2018), hu- tive and predictable (Surowiecki, 2017). This mans may soon develop their first zero pollu-

25 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018 tion, sustainable energy resource. This, along References with further advancements in AI, would heav- ily influence the industries including the man- 9 top jobs in 2030: Future skills you need ufacturing and agriculture sectors. This would to learn now. (n.d.). Retrieved eventually reduce global resource prices and November 5, 2018, from https:// reach a stage wherein everyone would have www.crimsoneducation.org/ie/ access to resources. The same holds true for blog/jobs-of-the-future global inflation, where we may observe stable Agiwal, M., Saxena, N., & Roy, A. (2018, and controlled prices. Mar). Towards connected living: 5g enabled internet of things (iot). IETE It is a social responsibility of humans to cre- Technical Review, 1–13. Retrieved ate a balance with ecology by which robots from http://dx.doi.org/10.1080/ backed up by AI would contribute to the hu- 02564602.2018.1444516 doi: 10 man community and its wellbeing; food gen- .1080/02564602.2018.1444516 eration, disease prevention, and medical care Alvarez, P. J. J., Chan, C. K., Elimelech, M., should be given the first priority. This would Halas, N. J., & Villagran,´ D. (2018, Aug). help us ensure a balance and mutual trust be- Emerging opportunities for nanotechnol- tween AI based technology and humans. ogy to enhance water security. Nature Nanotechnology, 13(8), 634–641. Re- For now, humans should focus on utilizing trieved from http://dx.doi.org/10 these intelligent platforms to solve global chal- .1038/s41565-018-0203-2 doi: 10 lenges such and global warming, climate .1038/s41565-018-0203-2 change, and resource management. Further- Bansal, P., & Kockelman, K. M. (2018, more, humans would need to be educated Mar 01). Are we ready to em- on the potential of such technological devel- brace connected and self-driving opments and refrain from exaggerating future vehicles? a case study of texans. outcomes to a point where it alters the ex- Transportation, 45(2), 641–675. Re- pectations of general public on our future with trieved from https://doi.org/ AI. As change is inevitable, humans would 10.1007/s11116-016-9745-z doi: learn to adapt and live with AI powered hu- 10.1007/s11116-016-9745-z manoids and robots in a harmonious manner. This would be accomplished through shed- Barone, R. (2017, December). 9 reasons ding our highly opinionated self and embrac- your child should learn to code (and one ing AI based technologies as our way forward word of caution). Retrieved November 5, www.idtech.com/blog/ into the future. 2018, from 5-reasons-your-child-should In summary, humans will continue to see fur- -learn-to-code ther advancements in the field of technol- Barrat, J. (2015). Our final invention: Arti- ogy and experience its impacts on their lives. ficial intelligence and the end of the hu- In these situations, humans are required to man era. St. Martin’s Griffin. change their mindsets, be informed of the Best, J. (n.d.). Ibm watson: The inside case, and use their ability to adapt to the en- story of how the jeopardy-winning vironment. Artificial Intelligence would play a supercomputer was born, and what prime role in future societies. However, de- it wants to do next. Retrieved spite widespread controversies, AI systems of November 5, 2018, from https:// the future would be cooperative and collabo- www.techrepublic.com/article/ rative with humans, providing them with the ibm-watson-the-inside-story necessary data and suggestions to speed up -of-how-the-jeopardy-winning the completion of job related tasks. This may -supercomputer-was-born-and hold true only if strict security measures are -what-it-wants-to-do-next/ taken to ensure that these intelligent systems Byttner, K. J. (2017, October). There’s only are not utilized for the wrong purposes. In con- one job ai can’t replace. Retrieved clusion, humans may look forward to an inter- November 4, 2018, from https:// esting future where adaptability is key for hu- nordic.businessinsider.com/ man evolution. theres-only-one-job-ai-cant

26 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

-take-over-2017-10 doi: 10.1007/978-3-319-90303-3 5 Chong, F. T. (2018). Quantum com- Knight, H. (2018, May). How humans puting is getting real. Proceedings respond to robots: Building public of the Twenty-Third International Con- policy through good design. Re- ference on Architectural Support for trieved November 5, 2018, from Programming Languages and Operat- https://www.brookings.edu/ ing Systems - ASPLOS ’18. Retrieved research/how-humans-respond from http://dx.doi.org/10.1145/ -to-robots-building-public 3173162.3177152 doi: 10.1145/ -policy-through-good-design/ 3173162.3177152 Kyle, S. (2018, Apr). 3d printing of Education - teaching code. (n.d.). bacteria: The next frontier in bio- Retrieved November 4, 2018, fabrication. Trends in Biotechnol- from https://www.apple.com/ ogy, 36(4), 340–341. Retrieved education/teaching-code/ from http://dx.doi.org/10.1016/ Frey, C. B., & Osborne, M. A. (2017). j.tibtech.2018.01.010 doi: 10 The future of employment: How sus- .1016/j.tibtech.2018.01.010 ceptible are jobs to computerisation? Lafrance, A. (2015, Mar). The 1980s: Technological Forecasting and Social An era of computerphobia. Re- Change, 114, 254 - 280. Retrieved trieved November 4, 2018, from from http://www.sciencedirect https://www.theatlantic.com/ .com/science/article/pii/ technology/archive/2015/03/ S0040162516302244 doi: https:// when-people-feared-computers/ doi.org/10.1016/j.techfore.2016.08.019 388919/ Gornall, J. (2014, September). Will we ever Lu, H., Li, Y., Chen, M., Kim, H., & Serikawa, understand the human brain? Retrieved S. (2018, Apr 01). Brain intelligence: Go November 4, 2018, from https:// beyond artificial intelligence. Mobile Net- www.weforum.org/agenda/2014/ works and Applications, 23(2), 368–375. 09/understanding-human-brain/ Retrieved from https://doi.org/10 Hassabis, D., Kumaran, D., Summerfield, .1007/s11036-017-0932-8 doi: 10 C., & Botvinick, M. (2017, Jul). .1007/s11036-017-0932-8 Neuroscience-inspired artificial intelli- McCarthy, J., Minsky, M., Rochester, N., & gence. Neuron, 95(2), 245–258. Re- Shannon, C. (1955). A proposal for the trieved from http://dx.doi.org/10 dartmouth summer research project on .1016/j.neuron.2017.06.011 doi: artificial intelligence. Retrieved Novem- 10.1016/j.neuron.2017.06.011 ber 4, 2018, from http://www-formal Hiner, J. (2018, January). Ai and jobs: Where .stanford.edu/jmc/history/ humans are better than algorithms, and dartmouth/dartmouth.html vice versa. Retrieved November 5, 2018, Panetta, K. (2018, August). 5 trends from https://www.zdnet.com/ emerge in the gartner hype cycle article/ai-and-jobs-where for emerging technologies, 2018. -humans-are-better-than Retrieved November 4, 2018, from -algorithms-and-vise-versa/ https://www.gartner.com/ Hype cycle research methodology. (n.d.). smarterwithgartner/5-trends Retrieved November 4, 2018, from -emerge-in-gartner-hype-cycle https://www.gartner.com/ -for-emerging-technologies en/research/methodologies/ -2018/ gartner-hype-cycle Pring, B. (2018, Feb). Why ai won’t take James, T. (2018). Scouting for resources. your job. Retrieved November 4, 2018, In T. James (Ed.), Deep space com- from https://www.cio.com.au/ modities: Exploration, production article/633045/why-ai-won-t and trading (pp. 69–79). Cham: -take-your-job/ Springer International Publishing. Re- Prox, J., Smith, T., Holl, C., Chehade, N., & trieved from https://doi.org/ Guo, L. (2018, Feb). Integrated biocir- 10.1007/978-3-319-90303-3 5 cuits: engineering functional multicellular

27 AI MATTERS, VOLUME 4, ISSUE 4 4(4) 2018

circuits and devices. Journal of Neural Wilson, T., Reinhard, D., Westgate, E., Gilbert, Engineering, 15(2), 023001. Retrieved D., Ellerbeck, N., Hahn, C., . . . Shaked, from http://dx.doi.org/10.1088/ A. (2014). Just think: The chal- 1741-2552/aaa906 doi: 10.1088/ lenges of the disengaged mind. Science, 1741-2552/aaa906 345(6192), 75–77. Rohani, M. M., & Pahazri, N. (2018, Apr). Sur- vey on how fluctuating petrol prices are affecting malaysian large city dwellers in changing their trip patterns. IOP Con- ference Series: Earth and Environmen- Sudarshan Sreeram tal Science, 140, 012085. Retrieved is a grade XII student from http://dx.doi.org/10.1088/ at Vidya Mandir Senior 1755-1315/140/1/012085 doi: 10 Secondary School, My- .1088/1755-1315/140/1/012085 lapore, Chennai, India. Russell, J. (2017, May). Google’s al- He is a dedicated stu- phago ai wins three-match series dent researcher and has against the world’s best go player. published two technical Retrieved November 5, 2018, from papers at international https://techcrunch.com/2017/ conferences: 18th IEEE 05/24/alphago-beats-planets EEEIC, Palermo, Italy -best-human-go-player-ke-jie/ and 2018 IEEE SMC Stanczyk, N. M., Brookfield, J. F. Y., Field, Junior, Miyazaki, Japan. L. M., & Logan, J. G. (2013, Feb). Aedes Sudarshan is an Apple aegypti mosquitoes exhibit decreased WWDC18 Scholar and the TUNZA Eco- repellency by deet following previous ex- Generation (Samsung-UNEP) Regional posure. PLoS ONE, 8(2), e54438. Re- Ambassador to India. His wish is to become a trieved from http://dx.doi.org/10 mature researcher in the future and contribute .1371/journal.pone.0054438 doi: to the field of computing. In addition, he is 10.1371/journal.pone.0054438 determined to advance his education to a Surowiecki, J. (2017, August). The great doctoral level. tech panic: Robots won’t take all our jobs. Retrieved November 6, 2018, from https://www.wired.com/2017/ 08/robots-will-not-take-your -job/ Valhouli, C. (n.d.). Cutting down the dis- sonance: The psychology of gullibility. Retrieved November 5, 2018, from http://www.columbia.edu/cu/ 21stC/issue-3.4/valhouli.html Whitehead, N. (2014, July). People would rather be electrically shocked than left alone with their thoughts. Retrieved November 4, 2018, from https:// www.sciencemag.org/news/2014/ 07/people-would-rather-be -electrically-shocked-left -alone-their-thoughts Why humans learn faster than ai-for now. (2018, March). Retrieved November 5, 2018, from https:// www.technologyreview.com/ s/610434/why-humans-learn -faster-than-ai-for-now/

28