Washington (CNN) Days after OpenAI CEO Sam Altman testified before Congress and proposed the creation of a new federal agency to regulate artificial intelligence, a US senator has introduced a bill to do just that.
On Thursday, Colorado Democratic Sen. Michael Bennet unveiled an updated version of legislation he introduced last year that would establish a Federal Digital Platform Commission.
The updated bill, which was reviewed by CNN, makes several changes to cover AI products more explicitly, including by amending the definition of a digital platform to include companies that “offer content that is primarily generated is through algorithmic processes.”
“There’s no reason the biggest tech companies on earth should have less regulation than Colorado’s small businesses — especially as we watch technology corrode our democracy and harm our children’s mental health with virtually no oversight,” Bennet said in a statement. “Technology is moving faster than Congress could ever hope to keep up with. We need an expert federal agency that can stand up for the American people and ensure that AI tools and digital platforms work in the public interest.”
The revised bill expands on the definition of an algorithmic process, clarifying that the proposed commission would have jurisdiction over the use of personal data to generate content or make a decision – two important applications associated with generative AI, the technology behind popular tools like OpenAI’s viral chatbot, ChatGPT.
And for the most important platforms – companies the bill calls “systemically important” – the bill would make requirements for algorithmic audits and public risk assessments of the damage their tools could cause.
The bill retains existing language mandating that the commission ensure platform algorithms are “fair, transparent and secure.” And under the bill, the commission would continue to have broad oversight authority over social media sites, search engines and other online platforms.
But the added emphasis on AI highlights how Congress is scrambling to make policy on an interesting technology it’s trying to understand. The debate over whether the US government should create a separate federal agency for police AI tools could become a key focus of those efforts after Altman’s testimony this week.
Altman suggested in a Senate hearing on Tuesday that such an agency could limit how AI is developed through licensing or credentialing for AI companies. Some lawmakers appeared receptive to the idea, with Louisiana Republican Senator John Kennedy even asking Altman if he would be open to serving as chair.
“I love my current job,” Altman said, to laughter from the audience.
Thursday’s bill does not explicitly provide for such a licensing program, although it directs the incoming committee to design rules appropriate for overseeing the industry, according to a Bennet aide. Bennet’s office did not consult with OpenAI on the original bill or Thursday’s amended version.
But even as some lawmakers have embraced the concept of a specialized regulator for Internet companies — which could conflict with existing police on the beat at agencies including the Justice Department and the Federal Trade Commission — others have warned of the potential risks of creating of a completely new bureaucracy.
Gary Marcus, a New York University professor and self-described critic of AI “hype,” told lawmakers at Tuesday’s hearing that a separate agency could fall victim to “regulatory capture,” a term that describes when the industry gain dominant influence over the government agencies created to hold them accountable.
Connecticut Democratic Sen. Richard Blumenthal, a former state attorney general who has prosecuted consumer protection cases, said no agency can be effective without proper support.
“I’ve been doing this stuff for a while,” Blumenthal said. “You can create 10 new agencies, but if you don’t give them the resources – and I’m not just talking about dollars, I’m talking about scientific expertise – (industry) will run circles around them.”
0 Comments