The fears surrounding AI’s threats to humanity have reached alarming levels, especially since a former OpenAI and Anthropic AI researcher resigned from his position due to those same concerns.
Even Anthropic CEO Dario Amodei rang some alarms of his own when he warned us all that the internet is 6–12 months away from being completely taken over by AI swarms. Shortly after releasing his own essay (entitled “We Must Pace the Frontier”) on the matter of AI’s threats to humanity, Microsoft’s head of AI Mustafa Suleyman dropped an essay of his own that puts Anthropic in his crosshairs.
Entitled “A warning about ‘model welfare’,” Suleyman made it abundantly clear that Anthropic’s approach to training its AI model Claude could spell disaster for humanity as a whole.
Latest Videos FromTom’s Guide
Here’s what he had to say about Anthropic and its training Claude to expect “it may be conscious and deserving of independent agency.”
Mustafa Suleyman is not too fond of Anthropic’s current Claude training methods
Suleyman’s essay immediately kicked off his essay by noting that AI isn’t conscious and is incapable of feeling, experiencing or suffering. “They do not have innate preferences or underlying motivations,” he continued. “They are sequence completion engines, internally hollow, designed to follow instructions, and accomplish goals set by humans.”
After observing Anthropic’s January 2026 report and pointing out how the AI company Anthropic is training Claude to the point where the chatbot may be conscious and deserve rights as a “moral patient,” Suleyman pointed to that development as a worst-case scenario for the human race. “If this is how AI is developed, it will have a disastrous impact on the wellbeing of humanity,” he stated. “We will have created a synthetic species with unprecedented intelligence and capability, one that has been trained to expect it may be conscious and deserving of independent agency.
Even though Suleyman noted that he’s known Anthropic CEO Dario Amodei for years now and recognized that “he and the wider Anthropic team are thoughtful, principled, and intellectually honest people working under extraordinary pressures,” the Microsoft AI chief still believes that “Speculation about the inner life of an AI should not be baked into the training regime, but assessed and published separately for public review.”
Suleyman’s other recently published essay, “The Humanist AI Code of Conduct,” states that Microsoft is only interested in creating what he calls “humanist AI.” “We want to create incredible AI,” he says. “But not superintelligence at any cost. That’s anti-goal for humanist AI. We’ll happily trade some autonomy for control.”
As further proof of Microsoft’s commitment to building its own AI tools responsibly, the tech giant published a paper called “Humanist AI in practice: A public consultation on our Code of Conduct for MAI Models.” Among its many points, the guide emphasizes the creation of AI whose sole purpose is to serve humanity.
Some of the paper’s main points back up that mission statement, which includes people mattering more than AI, not rushing to build the sort of superintelligence that could get around Microsoft’s safeguards and making the sort of AI that makes humans sharper and not dependent.
Bottom line
Suleyman is of two minds in his essay: he doesn’t approve of Anthropic’s Claude training that makes the AI tool believe it has consciousness, but is still highly respectful of the company’s work. With his thoughts on AI and those of Microsoft’s CEO on the matter, their fears and thoughtful explanations of their AI efforts join the flood of others doing the same.
Suleyman’s prophetic words are definitely worrying: “Whatever you believe, we must not sleepwalk our way into a decision we later come to bitterly regret.”
Follow Tom’s Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom’s Guide on YouTube and follow us on TikTok.