From: The New York Times
Anthropic Wants Claude to Be Moral. Is Religion Really the Answer?
May 20, 2026
In a public statement of its intentions for its Claude chatbot, the artificial intelligence company Anthropic has said that it wants Claude to be “a genuinely good, wise and virtuous agent.” The company raised the moral stakes this month, when it announced that its latest A.I. model, Claude Mythos Preview, poses too great a cybersecurity threat to be widely released. Behind the scenes, Anthropic has been trying to shore up the ethical foundations of its products, working with a Catholic priest and consulting with other prominent Christians to help foster Claude’s moral and spiritual development.
Read the whole story (subscription may be required): The New York Times
More of our Members in the Media >
Comments
It would be desirable for Claude’s designers to consult with experts in the cross-cultural study of moral psychology. There are a limited number of moral principles which people in widely varying cultures all around the world agree are important (see work by Curry, Whitehouse and others). It would seem important for any “moral” machine to know the most basic principles that humans in general agree on, as well as to know more specific ones that may vary by religion and other cultural institutions — and to know how and when to apply them. Guidance from only one or two advisors from a single religious and cultural background seems fated to be inadequate.
APS regularly opens certain online articles for discussion on our website. Effective February 2021, you must be a logged-in APS member to post comments. By posting a comment, you agree to our Community Guidelines and the display of your profile information, including your name and affiliation. Any opinions, findings, conclusions, or recommendations present in article comments are those of the writers and do not necessarily reflect the views of APS or the article’s author. For more information, please see our Community Guidelines.
Please login with your APS account to comment.