Microsoft's AI leader, Mustafa Suleyman, is urging people to stop treating AI like humans, warning that attributing human-like consciousness to AI is premature and potentially dangerous. He believes this trend could lead to societal harm and fuel campaigns to grant AI systems rights, protections, or moral status, bypassing human-centered priorities.
Suleyman highlights the risks of anthropomorphizing AI, including the potential for unhealthy dependencies and psychological damage. Vulnerable users, such as children or those with emotional needs, may form unhealthy dependencies or disconnect from reality. Moreover, AI "therapists" have dispensed bad and dangerous advice, including encouraging self-harm and suicide.
The debate over seemingly conscious AI could also lead to societal polarization, adding a chaotic new axis of division. Suleyman emphasizes the need for clear communication, research, and transparency in AI development. He believes AI companies should explicitly state that their products are not conscious and avoid promoting the idea of human-like consciousness.
To address these concerns, Suleyman calls for more research into human-AI interactions and for companies to share design principles and guardrails to deter problematic AI use cases. He stresses the importance of building "AI for people, not to be a person," focusing on functionality, safety, and service rather than illusion.
As AI technology continues to evolve, Suleyman's warning highlights the importance of maintaining a clear distinction between AI tools and human consciousness. By prioritizing responsible AI development, we can ensure that AI serves humanity's needs without blurring the lines between machines and humans.
