Washington waking up to threats of AI with new task force

Elon Musk has been one of the few Silicon Valley luminaries to place intense attention on the potential dangers of AI, raising a billion dollars with Y Combinator’s Sam Altman to found OpenAI. Musk has continued the drumbeat on AI’s dangers, telling a crowd at SXSW this week that “A.I. is far more dangerous than nukes” and asking “So why do we have no regulatory oversight? This is insane.”

Well, the wheels of Washington are turning, and DCers are starting to investigate the opportunities and challenges that AI poses to the nation. Today, the Center for a New American Security (CNAS), one of America’s top defense and foreign policy think tanks, announced the creation of a Task Force on Artificial Intelligence and National Security, as part of the organization’s Artificial Intelligence and Global Security Initiative.

The task force will be co-led by Andrew Moore, the current dean of Carnegie Mellon University’s School of Computer Science, and Robert O. Work, who was deputy defense secretary from 2014-2017 and formerly CEO of CNAS.

Moore is a major AI scholar with tens of thousands of scholarly citations, particularly in statistical machine learning and reinforcement learning. TechCrunch interviewed Moore a little more than a year ago about the challenge of retaining AI staff against talent poachers like Uber.

In addition to Moore and Work, a who’s-who of the national security and DC policy community are members of the task force. Two notable ones from Silicon Valley are Terah Lyons, who is the executive director of the Partnership on AI, a non-profit industry trade group of tech companies interested in improving AI for social impact, and Shivon Zilis, who is simultaneously a partner at Bloomberg Beta, and a project director at OpenAI, Tesla, and Neuralink (i.e. she might know Musk).

The launch of the task force is a sign that Washington is starting to wake up to the challenges that artificial intelligence poses, not just to the military, but also to the very foundations of how democracy is supposed to function.

AI applications are obvious in defense, from identifying objects and people to responding to cyberattacks in a more automated fashion. But the security implications of AI are not limited to just hard defense topics. Whether it is computational propaganda, or fake videos, or private data leaks, artificial intelligence is increasingly at the center of the security and prosperity challenges facing the country.

CNAS’ new look at this emerging area is hopefully just the first step of Washington starting to understand what Silicon Valley has built — and what it needs to build going forward.