Microsoft released a 37-page code of conduct for its AI model family on September 14, centered on the principle that humans come first. The document states that AI models are not conscious and should not be designed to imitate consciousness, while rejecting legal personhood for AI systems. The release follows Anthropic CEO Dario Amodei’s open letter warning about recursive self-improvement risks and calling for a coordinated industry slowdown, backed by Sam Altman and Elon Musk. The code mandates human oversight, transparency requirements, and explicit prohibitions on harmful applications, aligning with Microsoft’s Humanist Superintelligence approach and EU AI Act compliance. The broader safety conversation has coincided with a downturn in tech valuations, as Nasdaq-100 futures declined amid investor concerns over potential implications for aggressive AI growth projections.
Source: Read the original article
