Nadella Joins Call for Human Control in AI Development Amid Concerns of Unchecked Advancements

Microsoft CEO Satya Nadella has joined a growing debate over the need for human control in artificial intelligence development. The discussion was sparked by concerns that advanced AI systems could become difficult to control and pose risks to humanity.

Nadella, who is also Microsoft's chairman, expressed support for 'embedded evaluators,' which are third-party experts that continuously assess the safety of advanced AI systems. This approach has been proposed by other major tech figures, including OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei.

Nadella emphasized the importance of giving businesses greater control over AI development, allowing them to build their own continuous learning loops and embedding their knowledge into models without relying on a single provider. He also highlighted Microsoft's plans for a Code of Conduct for its first-party MAI models, which will be published for public consultation.

The debate has been fueled by warnings from experts about the potential risks associated with advanced AI, including loss of control, cyberattacks, and economic disruption. Amodei has proposed a three-part approach to address these concerns, including ongoing access to independent evaluators, coordination on common safety standards, and international cooperation.