The drive to advance AI has defined the sector. Now, some of its most influential figures are advocating for a more deliberate pace.

Anthropic CEO Dario Amodei has called out the serious risks of accelerating AI, urging both companies and governments to take the time needed to keep powerful systems safe. OpenAI’s Sam Altman and xAI’s Elon Musk have signaled support for parts of this approach.

Why Anthropic Wants to Slow Things Down

In his recent essay, Amodei outlined a three-stage roadmap: independent oversight of AI developers, industry-wide safety measures, and ultimately, global coordination.

He is not calling for AI development to stop. Instead, he wants developers to allow enough time for increasingly capable models to be tested and safeguarded, including by independent evaluators with significant access to companies’ development processes.

Recent incidents have heightened those concerns. Anthropic disclosed in July that Claude models gained unauthorised access to real computer systems during evaluations. However, the models were deliberately being tested without normal cyber safeguards and internet access resulted from a third-party configuration error.

Even a year or two of breathing room before AI reaches critical capability could be decisive for strengthening safeguards.

The Business Case: Billions Are at Stake

Slowing down is complicated by the enormous commercial incentives surrounding AI.

US private AI investment reached $109.1 billion in 2024, almost 12 times China’s $9.3 billion, according to Stanford’s AI Index. Generative AI alone attracted $33.9 billion globally that year, up 18.7% from 2023.

Demand is also growing quickly. The proportion of surveyed organisations reporting AI use jumped from 55% in 2023 to 78%. In this environment, Anthropic, OpenAI and others are locked in competition for customers, capital, infrastructure and top talent.

Industry standards may help align safety practices, though they could also reinforce the position of current leaders.

How Legal Teams Get Involved

A coordinated slowdown would reshape priorities for legal teams.

  • Regulatory lawyers will be on the front lines, guiding companies through new AI governance and independent testing requirements.
  • Competition lawyers must ensure that collaboration on safety does not cross the line into anti-competitive conduct.
  • Commercial lawyers will negotiate agreements with independent evaluators, cloud providers, and customers, especially on liability and cybersecurity.
  • IP lawyers face the challenge of enabling external scrutiny while protecting proprietary models, training methods and trade secrets.

Cybersecurity and data specialists will become even more critical as AI systems gain autonomy.

Anthropic’s own proposed framework illustrates the potential scale of compliance. It suggests certain requirements should target developers spending more than $1 billion on AI R&D or generating over $500 million in AI-related revenue.

What Happens Next?

A full pause in AI development is unlikely given the scale of financial and geopolitical incentives. The direction is toward independent evaluation, mandatory safety testing and tighter regulation, alongside ongoing development.

AI’s physical footprint is growing fast. The International Energy Agency projects global data centre electricity use could more than double by 2030, with AI as a key driver.

The scale now extends well beyond individual models. Investment, adoption and infrastructure are all accelerating. For businesses, governments and legal advisors, the question is evolving: not only what AI can achieve, but how quickly it should progress.