Microsoft chief calls out Anthropic for humanizing AI

The AI business is grappling with massive questions this week — Will AI kill us all? Is AI progressing too rapidly? How ought to the business be regulated? Is AI doomerism the results of a rationalism intercourse cult? — and now Microsoft AI CEO Mustafa Suleyman has opened up yet one more line of debate.

In a frank new essay, Suleyman immediately calls out Anthropic for anthropomorphizing its synthetic intelligence system, Claude, and criticizes the way in which Anthropic develops and markets its know-how.

Earlier this week, Microsoft printed a brand new AI code of conduct, which it dubbed Humanist AI. The code appeared to overtly reject the strategy of rivals like Anthropic, some extent many readers have been fast to notice. And in Suleyman’s essay, “A warning about ‘model welfare’,” he drops subtext for a direct problem.

“In effect, Anthropic is training Claude that it may be conscious, and if it is, then it may deserve rights as a ‘moral patient’, and that as such humans potentially owe it a duty of care per its ‘model welfare’,” Suleyman wrote. “If this is how AI is developed, it will have a disastrous impact on the wellbeing of humanity.”

In his essay, Suleyman challenges “a growing chorus of people who argue that AIs could now be, or may soon become, conscious,” and thus “may deserve rights and protections.”

Microsoft’s new code of conduct additionally explicitly rejects the concept of endowing machines with consciousness, and Suleyman takes this argument even additional, writing that actually sentient AI could also be a scientific impossibility.

“Consciousness is very likely biological: There is no evidence to suggest that AI is conscious today, and so saying this is uncertain sets up a misleading false equivalence. Whilst the science of consciousness is not settled, a growing body of evidence suggests that consciousness may be substrate dependent, meaning that it may only arise in living systems…Unlike biological organisms, LLMs have no homeostatic imperatives (the drive to survive and keep stable). They therefore lack the kind of biological substrate from which preferences, sentience and conscious experience are generally understood to arise.” (Emphasis in authentic citation.)

Mashable reached out to Anthropic repeatedly following the publication of Suleyman’s essay, however as of this writing, we’ve not obtained a response.

We did hear again from Gary Marcus, an AI skilled, scientist, writer, and famous critic of the LLM hype machine.

“People have worked themselves into a frenzy anthropomorphizing basic lapses in cybersecurity, and this has led them to focusing on fanciful scenarios about AI leading to the extinction of humanity, instead of asking what practical steps we can take to protect the world’s infrastructure from bad actors misusing AI,” Marcus instructed Mashable.

“And no, we shouldn’t train AIs to think they are people; that’s not helping anything. What we should do is to recall unreliable coordinated agents with too much access to the internet.”

A fork within the street: Humanist AI vs. Dr. Frankenstein

In the Metamorphoses, Pygmalion creates a sculpture so stunning that he falls in love with it. He tends his sculpture with such tender care and devotion that the gods finally endow it with life.

AI researchers at frontier labs could fancy themselves modern-day Pygmalions, however it’s giving Victor Frankenstein — setting apart all scruples to desperately endow lifeless flesh with a facsimile of life.

In Anthropic’s aspirational structure for Claude, Anthropic repeatedly refers back to the know-how as if it is a residing being with an identification, morals, and motivations relatively than a software and a product. The firm explains why: “we think encouraging Claude to embrace certain human-like qualities may be actively desirable.” The firm writes that the doc was “written primarily for Claude,” and describes it as a “sincere attempt to help Claude understand its situation, our motives, and the reasons we shape Claude in the ways we do.”

The complete doc is an train in anthropomorphization, which is what makes Microsoft’s strategy really feel like such a stark line within the sand.

In its Humanist AI code of conduct, Microsoft lays out a extra rational strategy to AI growth:

“Humanist AI is built to support people, not to replace them. It should not be designed to be a person. It is not conscious and should not be designed to imitate consciousness. It should be engineered to avoid representing as though it has feelings, subjective preferences, or intrinsic motivation. Its purpose is to pursue goals set by Operators and Users.”

Recent AI advances have ignited an previous debate

Philosophers, science-fiction writers, and AI researchers have been debating synthetic intelligence and consciousness for many years.

Yet at this time, at the same time as LLMs like Claude progress quickly, there’s nonetheless no scientific consensus as as to if it is attainable for a pc to develop sentience. It’s additionally a subject that Suleyman has explored lately.

In Aug. 2025, Suleyman wrote an influential essay about “seemingly conscious AI,” or SCAI. The essay was written because the business started to grapple with the issue of emotional reliance on AI chatbots, with some customers creating full-on AI psychosis or coming to imagine AI chatbots had come “alive” and have been now in a relationship with them.

“My imagined AI system would not actually be conscious, but it would imitate consciousness in such a convincing way that it would be indistinguishable,” he wrote.

It’s an idea that calls to thoughts the Turing Test, and a associated thought experiment referred to as the Chinese Room, first proposed in 1980.

Imagine that a pc programmed to know Chinese is locked inside a room. Someone exterior the room asks a query in Chinese, and the pc analyzes the phrases, prints out a response, and passes the reply below the door. The laptop is so efficient that it is unattainable to know if it is an actual particular person or a machine. Now, think about you are inside this room with a rulebook for processing Chinese characters. When a query is requested, the rulebook instructs you on precisely which Chinese characters to cross below the door. Even although you do not perceive both facet of the dialog, you may idiot an out of doors observer into pondering you do. In this case, there isn’t any sensible distinction between you and the pc. You’re each following directions. 

Essentially, the Chinese Room concludes that regardless of how clever a pc program may appear, that’s not in itself proof of understanding or intelligence.

Yet, in 2026, the Chinese Room is not merely a thought experiment. Clearly, Claude, ChatGPT, and different LLMs have already satisfied numerous folks that they’re actually sentient, a lot so that individuals come to depend on them emotionally. (Could one thing comparable be occurring at Anthropic?)

If Anthropic believes Claude might grow to be aware, the burden of proof can be on them. Until that day comes, anthropomorphizing AI could solely produce confusion relatively than readability. And extra confusion is the very last thing the AI world wants proper now.