Artificial Intelligence: Not Science Fiction, but Science Reality–Part 3

Lawmakers learned in April that the United States needs to write the script when it comes to artificial intelligence (AI). Leave this one to improvisation and we may have a tragedy on our hands.

The House Oversight and Government Reform Committee’s IT subcommittee on April 18 held its third and final hearing on AI to drill down on the appropriate role of the U.S. government in shepherding AI into the future.

With the technology outpacing legislation and even comprehension, government turned to key industry figures for guidance. The witnesses were asked if the current regulatory framework is sufficient to reel the technology in.

“I think this is a question of existential importance to basically the world,” said Jack Clark, director of OpenAI. “Because of the nature of the technology, traditional arms control regimes or other policy tools are insufficient.”

He said America is poised to be the thought leader in AI altruism, and needs to push for universal guidelines, or norms, on how technology should be applied.

“What I mean by norms are developing a global sense of what is right and wrong to do with this technology,” Clark said. “It’s about taking a leadership position on norm creation so that we can also influence how AI is developed worldwide.”

Think of it as a Geneva Convention on AI. These are cyber weapons, and thus need to be governed by international accords, established values, and repercussions for misuse.

Rep. Will Hurd, R-Texas, noted the dichotomy at play in these advancing systems. While AI brings advances in cyber defense, the potential for abuse rises in step. He pressed the witnesses on how government should measure, track, and manage advances in the technology.

Clark pointed to the Defense Advanced Research Projects Agency’s 2016 Cyber Grand Challenge as a major moment for AI, where autonomous systems competed to defend a network from cyber attack, without human intervention once the competition commenced.

By structuring future innovation “competitions” similarly and around defined end-goals, government can effectively benchmark innovation and track technological progress.

“Every single agency has data. It has intuitions of problems it’s going to encounter, and it has competitions that it can create to spur innovation,” Clark said.

Set a target and see who gets there fastest. It’s a model for progress that Gary Shapiro, president of the Consumer Technology Association, felt could yield tangible benefits outside the virtual.

“We need big goals in this area,” Shapiro said. “Dropping the death rate from automobiles down by half by a certain date would be a very admirable goal that everyone in this country could rally around.”

Recent