Tech Companies Turning to Religions to Make Artificial Intelligence Moral Face Criticism

Companies developing artificial intelligence technologies are contacting religions and faith leaders to help models distinguish between right and wrong. Firms like Anthropic and OpenAI aim to benefit from the accumulated wisdom of thousands of years of traditions.

Tech company executives state that they have been meeting with philosophers and religious leaders in order to develop morally good models against the destructive potential of AI. Anthropic co-founder Chris Olah stated that these questions cannot be answered in laboratories and that religions are traditions carrying thousands of years of accumulated wisdom. OpenAI CEO Sam Altman, on the other hand, argues that ChatGPT should reflect a weighted average of humanity's moral views.

Companies aim to create a broad pool of moral ideas without remaining tied to a specific religion. In a blog post published in May, Anthropic announced that it had held discussions with academics, clergy, philosophers, and ethics experts from over 15 religious and cross-cultural groups. Furthermore, giants such as Meta, Google, Amazon, and Microsoft are known to have maintained contacts with the Vatican and religious communities.

Experts' Criticisms and Concerns

Anthropologist Webb Keane notes that for decades academics have opposed the widespread understanding that views religion as a simple set of rules and principles detached from the practical lives of believers. It is emphasized that one must take into account how different traditions and worldviews of religions can contradict one another.

Meta Oversight Board member Paolo Carozza expresses concerns that unique characters might be lost or homogenized in the process of bringing together different knowledge and traditions. Anthropic, meanwhile, states that its work is not to fit models into a single tradition, but rather to gather diverse views on how good character is formed.