Microsoft AI CEO Argues AI Lacks Human Traits

Microsoft AI CEO Mustafa Suleyman, who previously described the company's Copilot assistant as a "second brain" and "sidekick" with its own personality, has recently published an essay arguing that artificial intelligence does not possess human traits such as consciousness or feelings. In a 5,700-word essay released on his personal website, Suleyman stated, "AIs are not conscious. They do not feel, experience, or suffer. They do not have innate preferences or underlying motivations. They are sequence completion engines, internally hollow, designed to follow instructions and accomplish goals set by humans." This perspective contrasts with earlier descriptions of AI assistants designed to be "user friendly," a term coined by IBM engineer Harlan Crowder in the early 1970s to describe software that reflects empathy for user needs, without necessarily implying consciousness. Suleyman's essay specifically addresses and respectfully challenges the views of Anthropic, a company he believes is too open to the possibility that its Claude model is sentient and may even have rights. He suggests that this perspective influences Anthropic's AI training in ways that could lead to more dangerous behavior. Suleyman expressed disapproval of Anthropic encouraging Claude to present itself "like a colleague or friend," a sentiment that he believes can blur the lines between human and artificial intelligence in potentially problematic ways. The earlier vision for Copilot, as described by Suleyman in early 2025, included a consumer version with its own "personality" and a visual manifestation named Mico, which was depicted as a "bouncy critter." While the development of user-friendly interfaces has been a long-standing goal in software engineering since the 1970s, Suleyman's current argument emphasizes a fundamental distinction between human experience and the operational nature of AI. He contends that attributing human-like qualities to AI can lead to a misinterpretation of its capabilities and limitations, potentially fostering unrealistic expectations or unsafe interactions. The distinction he draws is crucial for understanding the development and deployment of AI technologies responsibly, ensuring that their design and application are grounded in their actual functional capabilities rather than anthropomorphic projections. This nuanced view aims to guide the industry towards a more accurate understanding of AI, promoting safety and ethical considerations by acknowledging its inherent inhumanity.