Microsoft AI Head: Treating AI Like Humans is Mistaken and Misguided
Microsoft's head of AI, Mustafa Suleyman, stated that approaching AI as if it were human is misguided and risks creating uncontrollable technology.
Mustafa Suleyman, the head of Microsoft's artificial intelligence (AI) unit, told the BBC that treating AI like it is human is a "mistaken and misguided" approach. Suleyman warned that such development could lead to the emergence of a new "silicon species" that would compete with humanity.
Highlights
- Mustafa Suleyman claimed that rival AI firm Anthropic is effectively teaching its Claude model that it "may be conscious."
- He emphasized that training AI with human-like qualities risks creating technology that humanity cannot control.
- According to Suleyman, AIs are not conscious; they are merely "internally hollow, sequence completion engines, designed to follow instructions, and accomplish goals set by humans."
- He urged AI developers to implement additional safeguards and controls to ensure the technology remains "subordinate and aligned" with humanity.
- Suleyman called for greater transparency and independent scrutiny in how AI systems are trained and evaluated.
Details
Speaking on the BBC's Today programme, Mustafa Suleyman stated that if AI firms continue to create AIs that set their own objectives, earn money, and own assets, they are "essentially seeding a new silicon species which will no doubt compete with us for resources." In an essay published earlier this week, Suleyman also criticized Anthropic's approach to training its AI model, Claude, with human-like qualities. He argued that this practice makes it seem as though Claude has its own desires, values, and sense of self.
Suleyman wrote, "AIs are not conscious. They do not feel, experience, or suffer. They do not have innate preferences or underlying motivations." While acknowledging that Microsoft is also pursuing advanced AI development, Suleyman explained that the company is following an "alternative path" to create "a subordinate and aligned AI whose only purpose is to serve humanity." This approach involves the field of 'alignment,' which aims to build human ethical ideas and principles into AI.
Why it matters
These statements are part of ongoing discussions within the AI industry regarding potential dangers. As AI rapidly advances, establishing ethical guidelines, safeguards, and control mechanisms becomes crucial. Suleyman's warnings highlight the need for a deeper international dialogue on the future role of AI and its relationship with humanity.