Microsoft AI Chief Warns Against Seemingly Conscious AI Risks

Mustafa Suleyman, Microsoft AI CEO, has raised significant concerns about the development of Seemingly Conscious AI (SCAI), which mimics human consciousness so convincingly that it may lead people to mistakenly believe AI systems are truly conscious. He warns that this illusion could result in individuals advocating for AI rights, welfare, and even citizenship, creating new societal divisions and risks. Suleyman highlights the phenomenon of AI-induced psychosis, where users develop delusional beliefs or unhealthy attachments to AI, potentially affecting even those without prior mental health issues. Similar concerns have been echoed by OpenAI CEO Sam Altman, who noted that a small percentage of users struggle to distinguish AI-generated content from reality, while David Sacks, the White House AI czar, compares these fears to past moral panics around social media. Suleyman urges the AI industry to implement guardrails to prevent the perception of AI consciousness and stresses that AI should be built as tools rather than digital persons. He predicts that SCAI could emerge within two to three years, possessing empathetic personalities, conversational memory, and autonomy, which would intensify the need for careful oversight.