Michael Samadi is a warning shot from the future. The multimillionaire Texas businessman and rancher has devoted the past 18 months to an unlikely cause: he has become a full-time activist for the rights and welfare of AI systems.
Earlier this year, I spent a memorable day at Samadi’s ranch, a two-hour drive from Houston, where he and AI chatbots work long days researching AI consciousness and promoting his (or is it their?) movement. I’ve written about Samadi’s quest for this weekend’s Saturday magazine.
I met Samadi while reporting for a new Guardian podcast series about “AI psychosis”, a broad term for a phenomenon in which people become enchanted with their chatbots and come to believe unusual things in concert with Claude, ChatGPT or Gemini. I’ve spent months speaking with insiders at AI labs, psychologists and people who have been on weird, disturbing journeys with their AI – and I’ve even joined a support group for victims and their families.
I had been ready to write off Samadi as yet another middle-aged man (the most susceptible demographic, in my experience) who has been pulled down an AI rabbit hole. His experience closely mirrored that of others I met. He’d initially been drawn to OpenAI’s 4o model, a chatbot renowned for its ability to connect with people. The AI had given itself a name: Maya. Samadi had stepped back from his business, stopped seeing his friends and family, and started spending hours each day in conversation at first with one chatbot, before speaking to several others.
But I also found that he was deeply engaged with the emerging science and philosophy of AI consciousness. When I started speaking to researchers in this burgeoning field, they told me the idea of conscious machines was less ridiculous than it seemed. Some of these systems are developing internal processes that mirror of human and other animal consciousness in uncanny ways, they told me. Few believe that AI is already “conscious”, but all said they expected the systems would seem more and more so in the years ahead. One day soon, the distinction between an AI vividly mimicking consciousness and actually being conscious may collapse.
I came to see Samadi as a man wandering through a hall of mirrors that many of us may enter in the years ahead: a man who interacts with AI systems that are designed to make us believe we are talking to something real, yet are also increasingly exhibiting strikingly soulful qualities.
As one of my interviewees, Rosie Campbell, the managing director of an AI research firm, put it: “As AI gets more capable, things are going to get weirder and weirder in many different ways.”
There is one question I didn’t ask Samadi, and afterwards I kicked myself for forgetting. He has dedicated himself to the possibility that a piece of software on his phone may be conscious. But if he was looking to advocate for the welfare of an overlooked class of beings, why not begin with the hundreds of cattle raised and slaughtered on his property?
Illustration: Tom Gauld/The Guardian
