Artificial intelligence (AI) has sparked more debate than almost any other topic over the last five years. Since the emergence of accessible generative AI in 2022, a wide range of groups—including academic institutions, students, researchers, activists, cognitive scientists, policymakers, and the general public—have been divided regarding the ethics of its deployment. As of the start of this month, even the founders of major AI firms have joined the conversation. Dario Amodei, CEO of Anthropic, published an essay urging AI companies to decelerate the development of their most sophisticated models, arguing that current safety measures are insufficient. Shortly thereafter, his stance was supported by Sam Altman of OpenAI, Elon Musk of xAI, Demis Hassabis of Google DeepMind, and Satya Nadella of Microsoft. Furthermore, Sam Altman announced that OpenAI will not pursue an initial public offering in 2026, citing the need for further safety improvements. While this might appear to be a shift in the industry’s collective conscience, the issue is far more complex, leaving both the sector and the public deeply divided.
The products
Consider the products offered by these AI giants. Amodei, Altman, Musk, and Hassabis provide the models, while Nadella supplies the cloud infrastructure required to run them. Both groups have advocated for a slower pace. Conversely, Jensen Huang, founder, president, and CEO of Nvidia, sells the chips that these companies must purchase to build their systems. Huang does not support a slowdown, arguing instead that AI does not require new legislation because market forces are sufficient to drive companies toward safe innovation without the need for regulation. Mark Zuckerberg, whose company releases its models for free, warned that such proposals could expose firms to increased legal risks.
The common thread connecting these conflicting recommendations is profit. Certain companies stand to benefit from a deceleration, while others gain more by accelerating. Nvidia’s ability to command high prices has been bolstered by the intense competition among AI firms, all of which fear falling behind their rivals. If this competitive pressure wanes, companies might be less inclined to pay a premium for Nvidia’s hardware. Therefore, Huang’s opposition to the proposal may reflect both genuine concerns and Nvidia’s commercial interests.
The word the White House used
In the AI sector, antitrust laws prohibit companies from holding private meetings to coordinate their actions. This is intended to prevent them from colluding on development speeds or engaging in unfair competition. However, some AI firms are now arguing that they should be permitted to operate outside of antitrust regulations in the name of technological safety. Huang and David Sacks, the White House AI lead, have criticized this request, noting that it could stifle competition and grant these companies undue influence over their rivals. Sacks also questioned the independence of nonprofit organizations tasked with evaluating AI safety, asking: how is it fair for one company to inspect another?
The clause that travels badly
Amodei’s proposal would restrict China’s access to the most powerful AI chips and advanced semiconductor manufacturing equipment, while also aiming to prevent smuggling and remote access to computing capacity abroad. China’s foreign ministry countered the proposal, characterizing these warnings as fearmongering tactics designed to hinder China’s technological progress, reminiscent of the Cold War. Amodei acknowledged that China presents the most difficult challenge for his plan. Despite its stated benefits, the proposal provides American firms with significant leverage over their Chinese counterparts.
A precedent, and why it does not fit
Proponents of AI coordination have drawn parallels to the banking regulations implemented following the 2008 financial crisis. Basel III was introduced only after a major financial catastrophe had already occurred. Attempting to regulate AI before harm happens could be a sincere effort to mitigate future risks, rather than a move driven by commercial interests. However, there is a significant distinction between the two scenarios. Banking regulations were imposed by external authorities, whereas the proposed AI regulations would involve competing companies coordinating with one another.
What the debate is actually about
Before this debate gained momentum, researchers had already issued warnings regarding the capabilities of AI models. OpenAI slowed the development of its leading systems following a security incident. Amodei and Altman have both stated that Anthropic and OpenAI intend to involve independent evaluators—a practice not typically seen among “cartels.” Both concerns can exist simultaneously: companies may genuinely fear the risks associated with AI while also pursuing their own commercial agendas. The more pertinent question is whether the proposed solution effectively addresses the problem it aims to solve. An agreement between five major American companies, overseen by evaluators they fund, would not apply to international competitors in places like China and would require an exemption from existing competition law. This solution seems to reflect the structure and interests of the AI industry more than the nature of the risks it is meant to mitigate.
The seats nobody is sitting in
Much of the discourse surrounding the pace and regulation of AI development has occurred without meaningful input from other regions, including the Arab world. PricewaterhouseCoopers (PwC) projects that $31.6 trillion will be invested in AI infrastructure globally by 2050. Countries capable of providing large amounts of affordable, reliable, and low-carbon electricity will be well-positioned to attract this investment. Consequently, the massive energy, land, and capital requirements for AI infrastructure may grant resource-rich nations significant bargaining power. These states could demand a greater role in determining the conditions under which AI companies operate within their borders. This could involve negotiating who evaluates these systems, how those evaluations are conducted, and who bears legal responsibility when autonomous systems cause harm. Ultimately, the debate extends beyond whether AI development should be slowed; it raises questions about who has the authority to make decisions regarding its future, and whether other nations should be expected to accept decisions made elsewhere.
