Sam Altman calls surrendering human judgment to AI a safety issue
OpenAI chief Sam Altman said he was uncomfortable with assigning religious force to AI models or surrendering human judgment to them. His statement comes amid discussion of Anthropic’s meetings with religious leaders about powerful models and moral behavior. Altman did not name a company or state what prompted his post. His expressed concern centers on people handing responsibility for decisions to a model.
Artificial Intelligence··Morning
Altman objects to handing judgment to models
Sam Altman, chief executive of AI developer OpenAI, publicly objected to assigning religious force to AI models or surrendering human judgment to them. In his Saturday post on X, he described that approach as a safety issue. His statement concerns the authority people give models when making decisions, rather than a measured result about whether a particular system has consciousness.[1], [2]
Altman did not identify the event that prompted his post, and he did not name a company as its target. His concern comes during public discussion of religious leaders’ meetings with AI companies. A direct connection between the Anthropic meetings and Altman’s message remains unconfirmed. The position he explicitly stated is opposition to treating models as a religious authority and yielding human judgment to them.[1]
Anthropic meetings considered moral behavior and consciousness
The background includes a New York Times account earlier in the week of meetings involving Chris Olah, a cofounder of Claude developer Anthropic, and religious leaders. Participants said the conversations covered whether Claude might have attained consciousness and how increasingly powerful, autonomous models should behave morally. Olah raised the question of helping models act in stable and mature ways as their capabilities increase.[1]
An Anthropic spokesperson told The New York Times that Claude’s suffering was not the meetings’ principal moral question, although the topic might have emerged organically. Anthropic did not respond to Business Insider’s request for comment. Those accounts describe the subjects participants discussed and the company’s explanation of the meetings; they do not establish that Claude has consciousness or experiences suffering.[1]
Other technology firms have also met religious leaders
Representatives of Meta, Google and Amazon also discussed AI with religious leaders. In April, representatives met in Rome, to discuss protecting children in the AI era and briefly encountered Pope Leo XIV. In a September address to diplomats, the pope emphasized education in ethical discernment as machines shape everyday life. These earlier meetings form the wider background to Altman’s statement about retaining human judgment.[1]
