Jakarta, ThedailyID — Microsoft CEO Satya Nadella says advanced artificial intelligence (AI) systems need stronger safeguards, independent oversight, and emergency shutdown mechanisms. He believes authorities should be able to pause or deactivate AI models while they are performing tasks.
Nadella shared his views in a post on X on Saturday, October 10, 2026. He argued that increasingly sophisticated AI systems require reliable controls, especially when their behavior becomes difficult to predict.
“We need to equip models whose behavior is not always predictable with robust system design, human control, and reliable operating procedures,” Nadella wrote, as quoted by CNBC. He also called for industry-wide standards if existing rules prove insufficient.
Furthermore, Nadella suggested treating advanced AI models as potential insider threats. This approach should apply to both closed systems and models with open weights, he said.
The Microsoft chief also emphasized observability, a principle that allows people to monitor and understand a system’s actions and processes. According to Nadella, this requires several safeguards working together.
These include using multiple AI models, maintaining activity logs that humans can understand, and conducting continuous system testing. He also called for independent controls, audit capabilities, containment mechanisms, and incident reporting.
“The most trusted superintelligence systems are not those that use the models we trust the most. Rather, they are systems that require us to trust the models as little as possible,” Nadella said.
His comments come amid growing debate over the pace of AI development and the adequacy of existing safety measures. Technology leaders, including Bill Gates, Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and SpaceX CEO Elon Musk, have raised concerns about the technology’s potential risks.
Last month, an Anthropic AI researcher resigned and accused the company and OpenAI of putting human lives at risk through AI development. On the same day, an Anthropic AI alignment team leader warned that the technology could potentially kill all humans, estimating the possibility at more than 10%.
Meanwhile, US President Donald Trump has repeatedly dismissed concerns that AI could wipe out humanity. He has instead stressed the need for the United States to maintain its lead over China in AI development.
Trump recently introduced a task force called “AI Force,” led by US Director of National Intelligence Jay Clayton. The initiative aims to support the American AI industry while targeting the misuse of the technology.
As AI capabilities continue to advance, Nadella’s proposal highlights the growing emphasis on building systems that remain observable, controllable, and accountable.
