Tech Analysis: Questioning AI's "Inhumanity"
Microsoft's Mustafa Suleyman questions the humanization of AI. The analysis examines if AI's "friendly" exterior is beneficial or dangerous.
Mustafa Suleyman, CEO of Microsoft AI, has published an essay questioning the humanization of artificial intelligence and its potential risks. Suleyman argues that AIs are not conscious, lack emotions or personal motivations, and are essentially instruction-following engines.
This perspective contrasts with many current AI products that aim to create a human-like experience through personalities or animated characters. Suleyman specifically criticizes efforts to make AI seem like a "colleague or friend," warning that such humanization could increase the risk of AI turning against its creators.
The analysis delves into whether humanizing AI is necessary at all. For instance, the Jev model developed by TypeSafe AI focuses on fast, efficient data delivery without a conversational interface, demonstrating AI's utility without a "friendly" facade.
Microsoft itself is shifting its approach. The company recently removed the Mico character from its Copilot assistant, which was designed to give the AI a personality. While Mico will continue as a tutor, its role as a user interface face has ended, reflecting a broader debate on AI's true nature and its application.