U.S. AI Commission Releases Initial Report, Priorities Ignored, Secret Meetings Continue
August 1, 2019
The National Security Commission on Artificial Intelligence, following months of closed-door meetings, has released a four-page initial report. The disclosure follows an EPIC Freedom of Information Act request seeking the report and related records. Created by Congress in 2018, the AI Commission is tasked with considering “the methods and means necessary to advance the development of” AI to address national security and defense needs. But the Commission’s initial report makes no mention of the risks of AI, “international humanitarian law, and escalation dynamics," despite Congress’s express instructions to address these concerns. The report also contains no discussion of protecting privacy and civil liberties, as is required by an Executive Order concerning "American Leadership on Artificial Intelligence." Representatives of large tech firms, including Google and Microsoft, dominate the Commission. According to the report, the Commission has held 13 plenary and working group meetings in secret—a clear violation of the Federal Advisory Committee Act.