Sounding an AI alarm in Congress over China

In a report to be published this morning and shared first with Axios, the leaders of the House Oversight and Reform IT subcommittee — Chairman Will Hurd and Ranking Member Robin Kelly — call on the U.S. government to step it up:

"The United States cannot maintain its global leadership in AI absent political leadership from Congress and the Executive Branch."

The government hasn’t moved with the urgency the situation requires, Hurd, a Republican from Texas, told Axios. China’s rapid AI rise should shock Congress and the White House into action, he and Kelly write.

• Just yesterday, the White House convened a summit on quantum computing to work toward a strategy for supporting research in that much less developed field.
• The progress contrasted with stagnation in planning for AI, a technology that has been around much longer.

The pair of legislators lay out high stakes for failure. "Whoever masters AI will have an outsize role in this world," said Hurd.

• The nation with the strongest AI program will achieve a more efficient economy and improve decision-making in every industry. It will also have access to autonomous weapons, devastating cyberattacks, and supercharged disinformation and propaganda.
• And the first mover will get to set vital international norms and standards. "We should make sure how this topic is viewed around the world is based on free-market, western, liberal thinking," Hurd said.

The report offers four recommendations:

• More funding for R&D through agencies like the National Science Foundation, National Institutes of Health, DARPA, and NASA.
• Publishing government data sets, a potential boon for training data-hungry AI.
• Developing standards for measuring the progress and dangers of AI.
• More DARPA Grand Challenges, like the one that motivated much of the early work on autonomous driving.

Focused on broad themes, the report is thin on specifics.

• It calls for R&D funding, but doesn't say where money should go, said William Carter, deputy director of technology policy at the Center for Strategic and International Studies. He says funds should focus on basic research, often overlooked by the private sector, and on mitigating the potential societal harms of AI.
• The government should also promote the technologies that AI will depend on, like 5G and robotics, Carter wrote in an email to Axios.

While it's early to be implementing rules, some areas show promise for regulation, said Paul Scharre, director of the technology and national security program at the Center for a New American Security.

• Disclosure laws that would require notice when AI is mimicking a human — like Google’s Duplex, a virtual assistant that calls up businesses with a natural-sounding voice — is one, said Scharre.
• Deepfakes — computer-generated audio, video, and images — won’t be defanged with government regulations alone, he said, but raising public awareness about their danger is vital.
• "A completely hands-off approach to technology will lead to problems, just like we've seen in social media," said Scharre.

Earlier this month, DARPA announced it’s investing $2 billion for research into more flexible and powerful AI. That’s progress, Hurd and Kelly write, as was a May summit on AI at the White House Office of Science and Technology Policy.

What’s next: "There's not the level of interest and urgency and immediacy that we need from government right now," Scharre said, beyond a handful of interested lawmakers and White House staffers. "There is no national strategy."

• The executive branch needs to take the lead, he said, especially in driving funding.
• One immediate spark that could boost AI research: Opening vast government datasets, like information gathered from a vast network of weather sensors, said Hurd.