AI summary
Summary of article:
The technology industry wants to use religion and philosophy to make AI good, but scholars believe that the conflicts between different traditions are difficult to translate into universal rules; Anthropic collaborates with multi-religious groups to seek broader ethical context.
While technology companies compete to create the most powerful artificial intelligence (AI), some industry players are starting to think about how to teach chatbots to distinguish between right and wrong, and hope to use the “metaphysics at the end of science” to help “educate” AI.
Associated Press reports,anthropic Co-founder Chris Olah attended the event in early May this yearVatican publishedPope Pope Leo XIV said in his encyclical on AI: “These are not questions that laboratories can answer.”
As tech industry executives warn about the harms caused by AI and begin to discuss the need to create models of moral goodness, people in the tech industry are increasingly turning to philosophers and religious leaders for advice; A common view among industry executives is that religion may be able to guide AI to become good as fear of AI is spreading in society.
“I think ChatGPT should reflect a ‘weighted average’ of human morality or something like that,” OpenAI CEO Sam Altman said last year.
But experts and believers alike warn that this idea ignores the contradictions between different traditions and worldviews, which is not how religion works.
Webb Keane, professor of anthropology at the University of Michigan, points out that for decades, academics have resisted seeing religion as a set of rules that can be abstracted from the “disorganized, practical, adaptable” way of life of believers.
Keane believes that if religion can provide a set of extractable moral principles, then the training of AI could favor some traditions or distort others. “It’s a very old-fashioned idea to think of religion as a rule book,” he said.
This does not mean that the discussion lacks diversity. Meghan Markle, philosopher of religion and director of the ethics and AI program at the University of Notre Dame in Indiana. Meghan Sullivan attended an anthropological conference earlier this year that included Christians from various denominations.
Anthropic said in a blog post in May that it had met with scholars, clergy, philosophers and ethics scholars from at least 15 religious and intercultural groups and hoped to expand its scope in the future, but it did not list those it consulted.
Anthropic explained that the work is not to align the language model with any single traditional worldview, but rather hopes that the cloud will attract a wide range of religious, secular, and political viewpoints. “What we try to do through these conversations is to think critically and cumulatively about what constitutes good character.”
Representatives from Meta, Google and Amazon have also met with the Vatican, while Microsoft has been active in religious outreach.
Summary FAQ
-
As the industry is concerned about the harm caused by AI, it has started to think about how to equip chatbots with ethical judgment. When technology itself fails to answer questions of value, some companies turn to philosophers and religious leaders for guidance.
-
Experts believe that religion is not a set of fixed rules that can be directly derived and that there is often conflict between different traditions and worldviews. If religion is reduced to a rule book, AI training may favor certain beliefs and even distort other traditions.
-
Anthropic said it has communicated with at least 15 religious and cross-cultural groups, including scholars, clergy, philosophers and ethics scholars. The goal is not to align with any one faith, but to allow the cloud to assimilate diverse viewpoints such as religion, secularism, and politics.
