Microsoft CEO Nadella: We must Accept All AI Models Open to Danger

Microsoft CEO Satya Nadella attracted attention to the risks in the nature of artificial intelligence models in a long article released on the X platform. According to Nadella, now the era of acceptance of AI systems as a completely reliable black box has passed. We must assume that all advanced AI models are potentially dehlike or manipulatable.

The recommendations of Nadella include:

-Transparency request: Making AI models as “inner black boxes” instead of becoming open, traceable systems

• Delivery and control: Continuous monitoring and limiting the behavior of models
• Public registration: It is impossible to play on systems, leaving readable tracks by human
• Nötralization authority: Ability to disable suspicious AI decisions

Nadella’s call speeds up discussions on security and responsibility in AI industry. It seems that large tech companies such as Microsoft can not fully trust even their own models, they will increase their prints on AI regulation.

This approach shows that human control and control mechanisms are critical as the power of AI is increasing. AI systems used in the future should be traceable and responsible not only effective.