Bill Gates urges legally required AI safeguards as global investment tops $1 trillion

Bill Gates warned that AI businesses shouldn’t be allowed to self-regulate and is urging Washington to set safety regulations.

The Microsoft co-founder made his arguments during an NBC News interview released on September 25, two days prior to the airing of the complete interview on “Meet the Press.”

He has given this warning at the time when Goldman Sachs predicts global AI investments to go over 1 trillion dollars in 2026. This means that the new safety regulations may lead to important consequences for the developers, investors, and companies who are working on implementing the new security systems for the enforcement of the rules.

“No one thinks self-regulation is enough”

Gates responded: “Absolutely” when moderator Kristen Welker queried Congress’s need for action.

According to Gates, “No one thinks self-regulation is enough.” He believes that defined laws should be laid down by lawmakers and law enforcement bodies on how AI companies monitor risks and what safeguards they must put in place. He described the responsibilities on developers as “a little bit of overhead,” rather than something that slows progress.

His worry goes well beyond the usual misuse. According to Gates in an interview with NBC, AI is “certainly powerful enough to drive events that, you know, cause a billion deaths,” especially when these advanced systems fall in the hands of those with bad intentions.

Industry leaders have different opinions on how far those safeguards should go. Anthropic CEO Dario Amodei suggested the idea of “full pacing, or even ‘pause’” as the most extreme kind of collective slowdown. OpenAI CEO Sam Altman has expressed a similar concern, arguing that “pacing will be well worth this cost” and that the pressure of keeping ahead of the competition should not be an excuse for acting recklessly.

However, Meta CEO Mark Zuckerberg has said otherwise. In an interview with NBC News he stated, “I don’t think that we need some kind of industrywide coordination.” He prefers that labs slow down only when they feel they are at risk of safety issues.

Gates wants risks managed, not development halted

Gates is not only asking that the technology sector stop developing AI systems. In an essay published in his Gates Notes in August, he stated that artificial intelligence “will either be the greatest equalizer ever invented, or the worst source of injustice.”

He also remarked that he would likely endorse a credible plan for a global slowdown in AI development, though he was not convinced that such a plan could be effective because there are both economic and political reasons for continuing development.

His stance, thus, does not really consist merely of putting an end to new developments; it is, rather, about preparing companies, governments, and individuals while AI technology finds its application in such fields as medicine, law, software, and customer service.

A $1 trillion market waiting on the rules

The debate over safeguards is unfolding at the same time as money is pouring into AI, making the stakes much higher for any new rules. Goldman Sachs expects global AI investment to exceed $1 trillion in 2026, including about $581 billion in the US, while cumulative investment since 2022 could approach $1.8 trillion by year-end.

Despite this growth in AI investment, Microsoft says that not everyone has enjoyed equal access to its advantages and benefits. Microsoft’s AI diffusion data shows a growing divide between the Global North and Global South, while the OECD highlights that AI capacity is still limited to a handful of countries and firms.

This makes regulation an even more difficult balancing act. Better protections could decrease risks, but, for small labs and less wealthy countries, the additional cost of complying with these regulations could prove very hard to bear.

A regulatory patchwork is already forming

The call for stricter regulations is supported by a growing number of evidence. The International AI Safety Report 2026, written under the guidance of Turing Award winner Yoshua Bengio and created by more than 100 experts in the field, describes the existing dangers and the potential future threats posed by the increasing capabilities of AI systems.

Meanwhile, governments are already creating regulations around those risks. According to TRM Labs, a new global “patchwork” of AI regulation is emerging as different countries take different approaches to AI governance, implementation and accountability.

The European Union has pursued the implementation of such aims more aggressively than most other countries. According to the EU AI Act, providers of general-purpose AIs must comply with strict obligations in terms of documentation, copyright, and transparency regarding training data. Models utilizing more than 10^25 FLOP are seen to carry systemic risks, which corresponds to additional criteria for model evaluation, risk management, accident reporting, and cybersecurity.

With the increase in regulations, the amount spent by corporations attempting to comply with them is rising. According to Gartner, by 2027 the amount spent by organizations for securing AI will increase by 68.7% to almost $4.8 billion with projections to approach $7.7 billion in 2028.

Washington seems to be moving towards a more robust regulatory framework as well. As reported by Cryptopolitan, Senate negotiators are looking into a bipartisan initiative to establish ‘duty of care’ for AI developers which might allow the government to prevent the release of unsafe models. If this is approved, it will transfer the mechanism of protection that Gates is urging from the domain of voluntary obligations to legally binding responsibilities.