Truth, Lies, and Automation x How Language Models Could Change Disinformation AUTHORS MAY 2021 Ben Buchanan Micah Musser Andrew Lohn Katerina Sedova Established in January 2019, the Center for Security and Emerging Technology (CSET) at Georgetown’s Walsh School of Foreign Service is a research organization fo- cused on studying the security impacts of emerging tech- nologies, supporting academic work in security and tech- nology studies, and delivering nonpartisan analysis to the policy community. CSET aims to prepare a generation of policymakers, analysts, and diplomats to address the chal- lenges and opportunities of emerging technologies. CSET focuses on the effects of progress in artifi cial intelligence, advanced computing, and biotechnology. CSET.GEORGETOWN.EDU |
[email protected] 2 Center for Security and Emerging Technology MAY 2021 Truth, Lies, and Automation HOW LANGUAGE MODELS COULD CHANGE DISINFORMATION AUTHORS Ben Buchanan Andrew Lohn Micah Musser Katerina Sedova ACKNOWLEDGMENTS We would like to thank John Bansemer, Jack Clark, Teddy Collins, Aditi Joshi, Aurora Johnson, Miriam Matthews, Igor Mikolic-Torreira, Lisa Oguike, Girish Sastry, Thomas Rid, Chris Rohlf, Lynne Weil, and all the members of the CyberAI team at CSET for their comments on previous drafts and their help in preparing this study. In addition, we would in particular like to thank Catherine Aiken and James Dunham for their help in survey design and data collection, as well as Jennifer Melot for her assistance in data collection and code review. Lastly, we would also like to thank OpenAI for access to their model. All errors remain ours alone. DATA AND CODE Data and code supporting this project is available at https://github.com/georgetown-cset/GPT3-Disinformation AUTHORS Ben Buchanan is the director of the CyberAI Project and a senior faculty fellow with CSET, as well as an assistant teaching professor at Georgetown University’s Walsh School of Foreign Service.