With notable CEO’s in the AI industry sounding the alarm about rapid AI development, Online News Editor Beth Clifford explores the motivations behind the tech moguls’ proposals and how they might affect the global AI economy.

By Beth Clifford | Sep 28 2026

On Sept. 9, former Anthropic and OpenAI researcher, Jacob Coxen, took to X warning people that the global AI race poses a significant threat to society and civilisation, claiming that AI companies are “gambling with our lives” and they “earnestly believe that it [AI] could kill us all by the end of the decade”. Three days later, Anthropic’s CEO, Dario Amodei, published an essay proposing to slow the pace of AI development, acknowledging the technology poses significant harm if it remains unchecked.

This comes at a time when Anthropic is planning to go ahead with its initial public offering (IPO) that could value the company at $2 trillion. An IPO is when a private company offers shares to the public in a new stock issuance. Ownership transitions from private to public; this process is also known as a company “going public”. 

IPOs are generally sought out by a private company as a way to raise capital, pay off debts, fund new research, and invest in infrastructure. A private company may choose to go public as it can also improve its reputation, as it is required to disclose more information publicly, such as financial reports.

In his essay, Amodei suggests three steps to pace the rapid advancements in AI technology. The first step calls for the introduction of “Embedded Evaluators”, third-party regulators who are granted access, similar to that of an internal employee, to assess the work being carried out by frontier AI companies. Amodei states that these external regulatory bodies will receive desks, company laptops, access badges, and even contracts from the companies they are tasked with evaluating. 

Evaluators will be brought in to ensure AI companies are meeting safety standards and will be allowed to freely publish their findings to the public without interference from the companies unless it contains “security-sensitive, legally privileged, commercially sensitive, or third-party confidential information”. Evaluators are also permitted to clarify if a redaction by a company removes important information, allowing for greater transparency between the public and AI companies.

The second step outlined by Amodei is that frontier AI companies, most of which are American, ought to negotiate and collaborate with the US government to uphold a greater level of regulation through legislation. Amodei also proposes that the US government could mediate discussions between AI companies as an additional layer of state regulation to uphold safe practices in AI development. 

This step particularly highlights the strategic importance of keeping democracies’ lead in AI over autocracies such as China. Amodei believes achieving this could involve cutting off the trade of AI-manufactured inputs to autocracies and strengthening the software security of AI companies based in democracies. 

The third step seeks to coordinate states worldwide, which would see democracies and autocracies coming together to sign formal agreements in the hopes of slowing the pace of AI development. Defecting from such agreements to take the lead in global AI research and influence may seem appealing to superpowers such as China or the US; thus, Amodei considers that such formal arrangements should be made to emphasise the negative implications each party would face if any were to break from their promises.

Following the release of Amodei’s essay, CEO’s and scientists from leading AI companies came out and expressed their support for Amodei’s plans. CEO of OpenAI, Sam Altman, replied to Amodei’s X post with, “I agree with Dario that we need to pace the frontier”; Elon Musk of xAI also stated on X, “Dario is right”.

However, Amodei and fellow AI industry leaders have yet to clarify a complete timeline as to when AI development will be paced or come under increased regulation. As well as this, US President Donald Trump does not appear to be completely on board with AI companies regulating research, as he dismissed the threats as a “Hoax” on Truth Social.

The acknowledgement of AI's danger may hinder Anthropic’s financial prospects once its shares hit the stock market if investors believe the significant harms of AI will negatively impact their expected returns. Although, Amodei’s support for responsible AI research methods may calm the frenzy over AI's dangers and potentially boost Anthropic's reputation and its investments.

While American AI companies are preparing to pace their technological developments, the European Commission (EC) have recently revealed its plans to ramp up AI infrastructure. The EC is planning to build several AI gigafactories (AIGF) across the continent to boost AI sovereignty and increase the EU’s competitiveness in the AI industry, which is currently dominated by American and Chinese companies.

For scale, an AI factory is made up of approximately 25,000 AI-optimised chips, whereas an AIGF will involve up to 100,000 AI-optimised chips. The EC says that “€20 billion will be mobilised to finance the deployment of several AIGFs across the EU. This facility will be the largest public-private partnership in the world for the development of trustworthy AI”. Construction is set to begin sometime in 2027.