On August 26, Mustafa Solomon, head of Microsoft’s Artificial Intelligence Division, announced on social media that Booven was discussing the current state and future of artificial intelligence, this paper puts forward a developmental path of AI with“Human nature as the core”, and warns that society’s cognitive bias to the large language model (LLM) may lead to a deep ethical crisis.
Solomon stressed Microsoft’s current focus on turning AI tools, such as Copilot, into “Responsible Technologies” aimed at boosting human creativity rather than replacing it. The ideal AI, he argues, should have a“Human temperature”: one that is both deeply embedded in the real world and that promotes social trust and understanding through enhanced human collaboration.
“The real challenge is how to make AI work for humans without alienating them as cold tools,” Solomon writes, “Microsoft will be guided by the value of ‘putting people first’ to ensure that technological development is always in the service of human dignity and well-being.”
While envisioning the potential of AI, Solomon expressed strong concern about the current perception bias towards LLM. He points out that a growing number of people, including those without mental illness, are beginning to see the big language model as a conscious entity, even fighting for“Rights”, “Benefits” and even“Citizenship”. He defined the phenomenon as“AI Schizophrenia”– the over-anthropomorphizing of AI in human emotions while ignoring the ambivalence that it is essentially a probabilistic computing tool.
“When people are angry about CHATGPT’s ‘no answers’ or are depressed because of false information generated by AI, it’s not just a technical issue, it’s a psychosocial crisis,” warns Solomon, such cognitive dislocations can undermine human engagement in real relationships and even lead to the blurring of responsibility (for example, blaming the AI rather than the user for decision-making errors) .
According to the data, Mustafa Solomon, co-founder of DeepMind, joined Microsoft in 2022 to lead the AI department, leading the development of iconic projects such as Copilot and Phi-3. (Aoyama)