Jeff Bezos compared artificial intelligence and the data centers that sustain this technology to a knife, saying that banning these tools outright due to the risk of misuse is not the right approach. Bezos emphasized that the goal of regulation should not be to halt the technology entirely but to properly frame how it is used.
Bezos noted the important role of government regulations in public safety, citing aviation and pharmaceutical oversight as examples. He said a healthy regulatory approach can also be applied to new tools like artificial intelligence, but it should not be excessive.
Bezos Points to Regulation at the Application Level
Bezos, one of Amazon’s founders, expressed that regulation should focus on the "application level" of artificial intelligence. According to him, the issue is not targeting the technology wholesale but overseeing how it is used.
Explaining his approach, Bezos referred to the role of federal agencies in ensuring safety in areas like air travel and prescription drugs. Through these examples, he conveyed that new technologies can be managed without being completely unrestricted or entirely blocked.
High Computational Power for Prometheus
In addition to his AI advisory role at Amazon, Bezos serves as co-CEO alongside Vikram "Vik" Bajaj for Prometheus, a physical AI initiative. Bezos explained that he co-founded the company with Bajaj and that this is his first CEO role.
Prometheus has reportedly raised $12 billion in Series B funding. Bezos said the primary reason for such a large resource requirement is the extremely intensive computational power needed for the work being done.
Bezos also clarified that Prometheus is not focused on robot manufacturing. The company’s goal is to develop AI models that will transform engineering and manufacturing fields and ultimately provide tools that enable engineers to realize their ideas much faster.
Bajaj added that the initiative aims to develop a toolkit to help engineers turn their ideas into reality more quickly. According to Bezos, this approach is intended to move toward a kind of "artificial general engineer" concept.
Regulatory Debate Continues in Washington
Discussions about AI regulation have intensified over the past year. It is reported that companies are conducting intense lobbying against possible state-level regulations, while different oversight models are being discussed around the White House and Congress.
Former US President Donald Trump signed an executive order allowing AI model developers to voluntarily submit their models for federal review up to 30 days before public release. Trump also expressed expectations that leading AI companies commit to a form of "giving back" to the public.
Bezos did not directly respond to Trump’s order. Anthropic CEO Dario Amodei argued that models with computational power above a certain threshold should be subject to mandatory third-party testing by government or private organizations. Amodei wrote that the government should have the authority to halt a model if these tests reveal unacceptable risks.