Microsoft Chairman and CEO Satya Nadella called Saturday for advanced artificial intelligence systems to include containment measures, independent controls and an "emergency brake" that authorized personnel can use to pause or shut down a model mid-task.
"We need to surround non-deterministic models with strong, deterministic system design, human controls, and reliable operating procedures, and establish industry standards where existing ones are insufficient," Nadella wrote on X.
Nadella proposed treating frontier closed and open weight AI models "like insider risks" as a method to build safer systems. He outlined a set of design principles he called "observability," which include model diversity, human-readable records of model actions, continuous system testing, independent controls and auditability, containment and incident disclosure.
"The most trustworthy Super Intelligence system will not be the one with the model we trust most," Nadella wrote. "It will be the one that enables us to trust the model the least."
Nadella's statement adds Microsoft's voice to a growing chorus of concerns about AI safety. Microsoft co-founder Bill Gates, Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman and SpaceX CEO Elon Musk have all recently warned about insufficient AI safety protocols and claims that the technology is advancing faster than safety measures can keep pace.
Last month, an AI researcher departed Anthropic and publicly accused the company and rival OpenAI of "gambling with our lives." The same day, an alignment lead at Anthropic focused on AI safety stated there is a greater than 10 percent chance the technology could "kill all humans" within the next decade.
The debate over AI safety reflects a broader divide in how US leadership assesses the technology's risks. President Donald Trump has repeatedly dismissed AI extinction risks and instead emphasized the need for the industry to maintain technological advantage over China. Trump recently established a new "AI Force," led by Director of National Intelligence Jay Clayton, to facilitate industry development and identify bad actors.