Microsoft Chairman and CEO Satya Nadella called for advanced artificial intelligence systems to be built with containment, independent controls, and an emergency brake that allows authorized personnel to pause or shut down a model mid-task during a post published on the social media platform X.
Establishing Deterministic Controls
Nadella emphasized the necessity of surrounding non-deterministic models with robust system design and human oversight. “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. He added that treating both frontier closed and open weight models like insider risks represents a viable method for constructing such secure architectures. In his Saturday morning post on X, Nadella wrote that it is time to step back and assess the trust architecture
of AI, adding, We can’t treat Super Intelligence as a set of nested black boxes and simply accept or reject its recommendations, answers, and actions,
using the Trump administration’s preferred term for AI.
Expanding on these safeguards, Nadella described the approach as separating the model from the harness that orchestrates its work, alongside externalizing controls and safeguards. Furthermore, every meaningful model action must be documented with tamper-proof human-readable evidence. Nadella also stated, We must assume a model is compromised and contain it from the start.
Principles of Observability and Industry Warnings
Nadella outlined several core principles of observability required for these systems, including model diversity, continuous system testing, independent controls, auditability, containment, and incident disclosure. “The most trustworthy Super Intelligence system will not be the one with the model we trust most,” he wrote, “It will be the one that enables us to trust the model the least.”
These proposals emerge amid heightened industry warnings from prominent technology executives and researchers, including Microsoft co-founder Bill Gates, Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and SpaceX CEO Elon Musk, regarding insufficient safety protocols and the rapid pace of technological development. Commenting on the need for system containment, Nadella noted, Think of it like an emergency brake.
An AI researcher resigned from Anthropic last month, accusing the company and rival OpenAI of gambling with human lives. Later that same day, an alignment lead focused on AI safety at Anthropic stated there is a greater than 10 percent chance the technology could “kill all humans” within the decade.
Political Divergence and National Intelligence Oversight
In contrast to industry warnings regarding existential threats, President Donald Trump has repeatedly dismissed AI extinction risks, emphasizing instead the necessity for the domestic industry to maintain its competitive lead over China. To facilitate the sector and root out bad actors, Trump recently established a new “AI Force” led by Director of National Intelligence Jay Clayton.
