S Senator Proposes Bill to Establish Federal Agency for AI Regulation
Following OpenAI CEO Sam Altman’s recent testimony in Congress, where he suggested the creation of a federal agency to regulate artificial intelligence (AI), a US senator has introduced a bill to fulfill that purpose. Senator Michael Bennet, a Democrat from Colorado, unveiled an updated version of his legislation on Thursday, which seeks to establish a Federal Digital Platform Commission.
The revised bill specifically addresses AI products and introduces several modifications, including expanding the definition of a digital platform to encompass companies that primarily offer algorithmically-generated content. Senator Bennet emphasized the need for equitable regulation, stating that major tech companies should not be subject to less oversight than small businesses in Colorado. He expressed concerns about the impact of technology on democracy and the well-being of children, highlighting the urgency of establishing an expert federal agency to safeguard the interests of the American public and ensure responsible operation of AI tools and digital platforms.
The bill’s updates further clarify the scope of algorithmic processes by extending the commission’s jurisdiction to cover the use of personal data for content generation and decision-making—essential elements associated with generative AI technologies like OpenAI’s popular chatbot, ChatGPT. For platforms deemed systemically important, the bill introduces requirements for algorithmic audits and public risk assessments to evaluate potential harms caused by their tools.
While the bill retains its existing provisions mandating fair, transparent, and safe platform algorithms, it also reinforces the commission’s broad oversight authority over social media sites, search engines, and other online platforms. The increased focus on AI in the legislation reflects Congress’s growing efforts to navigate the policymaking landscape for this cutting-edge technology, which remains a subject of rapid exploration and understanding.
During his Senate hearing, Altman proposed the notion of a separate federal agency capable of regulating AI development through licensing or credentialing for AI companies. Some lawmakers expressed openness to the idea, with Senator John Kennedy of Louisiana even inquiring if Altman would consider serving as the agency’s chair, prompting a lighthearted response. Although the bill introduced on Thursday does not explicitly include a licensing program, it empowers the prospective commission to design appropriate rules for overseeing the industry.
As discussions unfold regarding the establishment of a specialized regulator for internet companies, potential concerns have emerged regarding the creation of an entirely new bureaucracy. Critics, including Professor Gary Marcus from New York University, have cautioned against regulatory capture, wherein industries exert disproportionate influence over government agencies tasked with holding them accountable. Senator Richard Blumenthal, a Democrat from Connecticut and former state attorney general, stressed that effective agencies require adequate support, including not just financial resources but also scientific expertise, to prevent industry dominance.
While the proposed bill represents a significant step towards regulating AI, the path forward will involve navigating various perspectives and ensuring the agency has the necessary resources and expertise to fulfill its regulatory responsibilities effectively.