Microsoft CEO Satya Nadella has raised concerns about strict control over Anthropic’s latest AI model, Fable. During an internal meeting with engineers working on Microsoft’s Copilot platform, Nadella discussed cases where Fable refused prompts that users considered harmless. He questioned whether creative AI systems should face such strong restrictions while responding to everyday requests.
Anthropic launched Fable 5 in June with updated safety measures designed to reduce harmful responses and limit incorrect refusals. The company later admitted that some genuine requests could still be blocked under the new safeguards.
The issue sparked debate among AI users and developers. Critics have questioned whether safety systems should prevent access to useful information in areas such as software development.
Nadella also argued that AI computing resources should not be under the control of a few companies. He said businesses should have more freedom to develop custom AI systems using their own data.
“It can't be that there are only two companies in the world with token capital, and everybody else is renting it. It makes no economic sense,” Nadella told engineers while discussing access to AI infrastructure.
Microsoft’s Azure AI Foundry platform currently provides access to more than 11,000 AI models. These include systems from Anthropic, OpenAI, and other developers, along with Microsoft’s own models.
Microsoft has invested heavily in AI through partnerships with OpenAI and Anthropic while also building its own AI technology. The company has expanded Copilot across the workplace and consumer products.
Nadella also stated that Microsoft should have combined its consumer and enterprise Copilot efforts earlier. The company has merged the leadership of both areas to strengthen its AI strategy.
The CEO’s comments highlight the shift in the AI industry. Experts have moved beyond model performance and are raising questions around safety controls, accessibility, and ownership of computing resources.
Also Read: China’s Backdoor Warning Puts Anthropic’s Claude Code Under Global Security Spotlight