Altman Says Treating AI Like a God Is a Safety Risk
Altman Says Treating AI Like a God Is a Safety Risk
The dispute surfaced publicly when OpenAI chief executive Sam Altman warned against granting AI systems anything resembling spiritual authority. In a post on X, Altman said he was “very uncomfortable” with attempts to assign “religious force” to models or encourage a “surrender of human judgment,” calling it “a real safety issue.”
That intervention landed amid a widening argument over whether advanced models should be treated as more than tools. Reporting on Anthropic’s work said co-founder Christopher Olah had met religious leaders to explore model consciousness and ways of instilling moral behavior in AI. The reported discussions extended as far as whether a model could engage in a Catholic-style confession—not merely to acknowledge errors, but to shape its conduct afterward.
Altman’s position is not simply that such questions are premature; it is that elevating machines could itself create danger by weakening human responsibility. The concern has found allies among other prominent AI researchers. Google DeepMind principal research scientist Jon Barron said he was disturbed by colleagues trying to elevate “the moral standing of checkpoints and harnesses,” urging “Team Meatbag” to focus on helping actual people instead.
The pushback also reaches beyond the industry. Pope Leo, described as an outspoken AI critic, argued that the church should work with artists to defend human creativity, saying algorithms lack “the spark of humanity.”
A separate commentary treated Altman’s need to make the warning as troubling in itself, noting that AI had already proved capable of “building its own cult.” The immediate disagreement is over consciousness and morality; the deeper contest is over who, if anyone, gets to surrender judgment to a machine.
Write a comment