Microsoft CEO Satya Nadella has joined industry leaders in calling for a more cautious approach to artificial intelligence development. While not advocating for a complete halt, Nadella emphasized that the pursuit of superintelligence must prioritize human control and benefit, stating that technology failing these tests is not worth developing. He supports deliberate pacing and the implementation of independent evaluators to ensure safety principles are effectively enforced.
Beyond safety, Nadella argued that the AI ecosystem must remain diverse and not be controlled by a small group of entities. He expressed concern over businesses becoming overly dependent on a few providers, suggesting that organizations should retain control over their own data and learning loops. To address these issues, Microsoft plans to publish a Code of Conduct for its first-party AI models for public consultation.
This intervention follows recent warnings from Anthropic CEO Dario Amodei, who proposed giving third-party evaluators access to AI labs to monitor safety compliance. OpenAI CEO Sam Altman and Elon Musk have also expressed support for such measures. The growing consensus among industry leaders reflects increasing pressure to address the potential risks associated with the rapid advancement of AI systems.
