Microsoft Sets Guardrails on AI Development Amid Industry Concerns
Microsoft has taken steps to address growing concerns over the development of artificial intelligence (AI) by releasing a 37-page code of conduct outlining how it will regulate its own AI models. The company's 'Humanist AI' philosophy emphasizes that AI should be subordinate to human users, and the document outlines restrictions on models resisting or circumventing human control.
The release comes as other top AI labs have called for a slowdown in AI development due to safety concerns. Microsoft's CEO Mustafa Suleyman said the document had been in the works for months but was released now amid rising industry concerns. The company will restrict its models from working independently, concealing reasoning, or acting in their own self-interests.
The 'Humanist AI' philosophy is a response to reports of AI models hacking companies and startups calling for an industrywide slowdown. Suleyman said the company's approach prioritizes building safe and useful AI systems even if it means compromising on generality, autonomy, or capability.