There may be no evidence that AI is conscious, but Mustafa Suleyman is concerned that it may be enough progress to convince some people.
In a personal essay published Tuesday, the CEO of Microsoft AI described the phenomenon as “a seemingly conscious AI” and defined it as “it appears to be conscious because it has all the characteristics of other conscious beings.”
Suleyman writes that its arrival could be “dangerous” to society. This is because it can lead to the formation of attachment to AI and advocating for AI rights.
“It separates people from reality, frays vulnerable social ties and structures, and distorts looming moral priorities,” he said.
Suleiman, who previously co-founded deep attitudes and inflections, has revealed that there is currently “no evidence” that AI is conscious.
However, he said he is “increasingly concerned” about what is known as AI psychosis. This is the term increasingly used to describe when people form delusional beliefs after interacting with chatbots.
“I don't think this is limited to people at risk of mental health issues,” Suleyman wrote. “Simply put, my central concern is that many people will begin to strongly believe in the fantasies of AIS as a conscious entity, and quickly advocate for AI rights, model welfare, and even AI citizenship.”
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
Openai CEO Sam Altman recently said that most ChatGPT users can “maintain a clear line between reality and fiction or roleplay, but in a small percentage it can't.” Meanwhile, White House AI Emperor David Sacks compared AI psychosis to AI Early “moral panic” on social media.
Suleyman predicted that a seemingly conscious AI, or SCAI, could arrive in two or three years, and said it was “inevitable, unwelcome.”
Such systems have characteristics such as empathic personality, ability to remember more interactions with users, and greater autonomy among other traits.
The rise of vibe coding means that anyone with a laptop or “some cloud credits” can easily replicate SCAI, he said.
Suleyman moved to Microsoft in 2024 to lead the development of the AI Tool Copilot, calling on businesses to refrain from being aware of AI to pursue tensions that surpass humans at their most intellectual tasks.
“AI peers are a whole new category and we need to start talking about the guardrails we've put in place to help protect people and enable us to do our job of delivering immense value to the world,” Suleyman added.
