Summary
- The Cato Institute cautions that a government-mandated halt on AI could stifle innovation and benefit established companies.
- Jack Dorsey shares concerns over negotiated limits by major developers while supporting some restrictions to mitigate catastrophic risks.
- Amid calls for regulation from lawmakers and AI leaders, analysts are skeptical about the effectiveness of voluntary measures.
The Cato Institute, a libertarian think tank, has issued a warning that a government-imposed pause on artificial intelligence advancement could provide an advantage to dominant firms while delaying the introduction of beneficial technologies. This caution comes as lawmakers and AI executives advocate for regulatory measures.
In a blog entry published on Monday, Jennifer Huddleston, a scholar in technology policy at Cato, suggested that companies could implement voluntary safeguards and establish shared standards to mitigate specific risks without halting industry-wide development.
Myriad: How low will Nvidia go? Click to make your prediction.“A government-mandated pause would raise multiple concerns and is likely to fall short of achieving its intended safety enhancements,” she stated, noting that governmental regulations are slow to formulate and even slower to adapt. “A regulatory framework designed for current models may obstruct future innovations, hindering potentially superior or safer solutions.”
Huddleston recognizes the risks associated with AI but cautions that regulations tailored to favor leading developers could enable them to solidify their market dominance, thereby limiting competition—an instance of market capture that would disadvantage smaller firms. She also warns that halting U.S. AI advancements could compromise cybersecurity as foreign competitors continue to progress.
Her remarks come in response to a proposal from Sen. Bernie Sanders and Rep. Greg Casar advocating for a pause on advanced AI development until federal safety standards are established, alongside a permanent ban on artificial superintelligence. Meanwhile, OpenAI CEO Sam Altman has expressed support for a more gradual development approach while urging companies to enhance safety measures without waiting for federal guidelines.
“Many of these discussions are often accompanied by science fiction-like doomsday scenarios about potential disasters,” Huddleston remarked. “However, a hasty regulatory response driven by fear could yield longer-lasting negative ramifications than the risks it aims to mitigate.”
Jack Dorsey, co-founder and chairman of Block, echoed similar sentiments regarding competition in a post on X on Monday.
“The companies spearheading machine intelligence should have a voice. They possess the expertise and commercial stakes to safeguard,” Dorsey stated. “Regulations centered around their resources could limit participation to only those firms. Genuine safety concerns could inadvertently create barriers to entry.”
Dorsey advocates for open releases that allow researchers to examine, modify, and test AI models, with evaluations and known limitations made public so that outsiders can challenge the safety assertions of developers.
“I support scrutiny but oppose industry-wide limits agreed upon by current leaders, as they could exclude individuals who might identify flaws or create alternatives,” he expressed. “Safeguarding a company's competitive edge is not synonymous with ensuring safety.”
In his statement, Dorsey also endorsed independent testing and accountability measures, including the possibility of withholding models if evidence suggests their release would significantly elevate catastrophic risks that cannot be sufficiently addressed through narrower strategies.
“Computational scrutiny can occur without stifling development. Examination should not require permission from a regulator or competitor,” he clarified. “I reject general approval prerequisites or waiting periods. Any imposed safety-related delay in publication must be justified by the catastrophic-risk exception.”
A recent analysis by the Atlantic Council posits that competitive pressures necessitate enforceable AI safety standards, while U.S.–China tensions complicate the formation of international agreements. Moreover, collaboration raises antitrust issues, as OpenAI has inquired whether rival developers could legally coordinate to decelerate development.