Microsoft AI chief criticizes Anthropic's approach to AI consciousness

Anthropic: Anthropic is an AI research company focused on developing safe and aligned large language models, primarily through its Claude family of systems. It incorporates detailed constitutional principles into model training to guide behavior and address potential risks. Its approach of including language about possible model consciousness and welfare in training materials has sparked public debate with other industry leaders over long-term control implications. Microsoft: Microsoft is a major technology company actively developing and deploying artificial intelligence systems across its products and cloud services. Its AI efforts emphasize safety, alignment with human goals, and maintaining human oversight of advanced models. The company's AI chief, Mustafa Suleyman, recently published an essay criticizing how competitors handle concepts of model consciousness in training data. AI Safety Debate: Microsoft's AI leadership has argued against embedding speculation about consciousness or moral status in AI training to avoid complicating human control over advanced systems. Industry Response: Anthropic continues to prioritize thoughtful safety research while advancing its models, as evidenced by its recent threat intelligence reporting on misuse prevention. Corporate Guidelines: Microsoft released a draft Humanist AI Code of Conduct emphasizing that AI models should remain subordinate tools serving humanity without claims to personhood.