i Origins February 24 — 26, 2017 PROJECT An Origins Project Scientific Workshop ARIZONA STATE UNIVERSITY Challenges of Artificial Intelligence: Envisioning and Addressing Adverse Outcomes e Noconsensus reality, inability to coordinate large-scale positive action e Concrete version of emergent social failure from Al technology DISCUSSION Consider the adverse outcomes with information flows and associated threats to democracy and freedom. What surprises might lurk in our future around costly outcomes in this realm? How might we thwart attacks on manipulating content from people, and on harnessing or hacking someone’s identity? What might be done to thwart a march to adverse outcomes for information, freedom of thought, democracy? What recommendations might be made about steps for moving forward? POTENTIAL GOALS e Seek a better understanding of the technological, social, political and economic aspects around uses of Al for generating, optimizing information and propaganda. e Identify potential blueprints for institutional interventions that may prevent/slow/detect the scenario unfolding e Develop ideas for coordinating relevant actors (advertising agencies, political parties) and/or carriers (media outlets, digital platforms) to prevent the worst versions of the scenario. e Identify potential approaches to thwarting attacks harnessing identify, including certification of identity by owners, identifying mechanisms for thwarting generation and distribution of false content. Possibilities of new approaches to minimizing threat with fines, other regulatory activity. REFERENCES The Secret Agenda of a Facebook Quiz, New York Times, Nov. 19 2016. https://www.nytimes.com/2016/11/20/opinion/the-secret-agenda-of-a-facebook-quiz.html?_r=0 Trump’s plan for a comeback includes building a ‘psychographic’ profile of every voter, Washington Post, October 27 2016. https://www.washingtonpost.com/politics/trumps-plan-for-a-comeback- includes-building-a-psychographic-pro