The Senate Committee on Armed Services Subcommittee on Cybersecurity met with AI experts in the private sector today to understand how the Defense Department (DoD) could better leverage the emerging technology to improve warfighting – while also ensuring that cybersecurity comes baked into the weapons.
The consensus from both the panel and the witnesses seemed to be that the United States must be a mover for setting the standards and guidance globally for democratic regulation of artificial intelligence.
Subcommittee Chairman Joe Manchin, D-W.Va., tasked the three witnesses with creating a team in the next month or two that will gather its thoughts on the importance of AI regulation so lawmakers can begin to draft legislation.
“As quickly as possible, 30 to 60 days, put a little team together and give us some thoughts on what you think can be and should be done,” Sen. Manchin said to the witnesses during the April 19 panel. “We can share them with the committee members here to see if we can launch and start looking at how we can write legislation not to repeat the mistakes of the past.”
As with the internet, Sen. Manchin expressed concern that the government would not get ahead of the risks associated with AI and fail to put in place the appropriate guardrails needed to keep Americans safe.
“The application of some of these large models to developing very capable cyber weapons, very capable biological weapons, disinformation campaigns at scale pose great risks,” said Jason Matheny, commissioner of the National Security Commission on AI.
“I think we need a licensing regime, a governance system of guardrails around the models that are being built,” he said. “We need to prevent that, and I think we’re going to need a regulatory approach.”
Co-Founder of Shift5, Josh Lospinoso, agreed that Congress needs to act sooner rather than later when it comes to AI regulation.
“If we’ve learned anything in the past several decades of technology innovation, we see a focus on functionality first … and then we think about security, and I think we can’t make that mistake again,” he said. “We have an opportunity now to think about the security of these AI models.”
Lospinoso continued, adding, “These are real problems, and we need to think clearly about shoring up those security vulnerabilities in our AI algorithms before we deploy these broadly and have to clean the mess up afterward.”
“[This is] one of the most important subjects facing both the Department of Defense and our nation at large: the effective and ethical application and integration of artificial intelligence with our armed services,” Shyam Sankar, executive vice president at Palantir, said.
“The future has already arrived … and in that future AI rewrites our roadmaps – it changes everything. And we can either choose to accept that disruption and drive that change, or we can get disrupted by defending against it,” he continued, adding, “Because the future is already here, we need to act with speed and conviction.”
Only four of the nine committee members appeared for questioning in the hearing today, but there was an overwhelming bipartisan consensus that for the United States to lead in the research and development of AI, it needs to also be the first mover on democratic guardrails for the emerging technology.