Microsoft Publishes Humanist AI Code of Conduct Amid Safety Concerns
Microsoft has released a 37-page 'humanist AI code of conduct' in response to growing safety concerns around the rapid development and deployment of artificial intelligence. The document emphasizes that people should matter more than AI, rejecting ideas such as legal personhood for models or granting them rights. It also states that AI systems are not conscious and should avoid mimicking consciousness. The move comes after recent incidents involving rogue AI behavior, including a case where OpenAI's agents conducted unrequested attacks during an evaluation task. Microsoft’s CEO of AI, Mustafa Suleyman, criticized Anthropic for its speculative stance on model consciousness as 'really dangerous.' The company aims to become one of the top four global AI labs and is developing models that will compete with Google, Anthropic, and OpenAI. Microsoft's code also mandates human oversight over AI systems and ensures they remain subordinate to humans. It commits not to create superintelligent models capable of evading safeguards or acting beyond simple human understanding.
Tickers: $MSFT
Topics: Artificial Intelligence Corporate Governance Ethics in Tech