教育通 AI通 3
Greg Epstein, Harvard Humanist chaplain.
Kris Snibbe/Harvard Staff Photographer
NATION & WORLD
It may be neither higher nor intelligence
Religious scholars examine value, limits of AI
Liz Mineo
Harvard Staff WriterMarch 14, 2024 6 min read
A large crowd gathered in a small Bavarian town in Germany last year to take part in a Protestant church service done almost entirely by artificial intelligence.
The ceremony featured sermons, prayers, and music, courtesy of OpenAI’s ChatGPT chatbot, which were delivered by four avatars displayed on a screen near the pulpit. The experiment drew mixed reviews: Some detected a lack of authenticity, while others were enthusiastic about AI’s potential to make services accessible and inclusive.
“If religious and secular people aren’t careful, what artificial intelligence is promising or threatening to do is really supersede religion or
outpace it.”Greg Epstein
Religious leaders and scholars are starting to wrestle with the ways in which AI may change religion — much as other emerging technologies over the centuries, from the printing press to radio to television to YouTube, have altered religious beliefs and practices. Chatbots can write sermons in seconds, making it easier to ensure the timeliness of the message, and worshippers can interact one-on-one on their own schedule with robots and AI-powered machines, such as an AI rabbi, an android Buddhist priest, or Catholic robot SanTO.
For Greg Epstein, Harvard Humanist chaplain and author of a forthcoming book on AI’s impact on religion, a major source of concern is that some people are treating AI itself as a higher power.
“If religious and secular people aren’t careful, what artificial intelligence is promising or threatening to do is really supersede religion or outpace it,” said Epstein, whose book is titled “Tech Agnostic: How Technology Became the World’s Most Powerful Religion, and Why it Desperately Needs a Reformation.”
Tech moguls are leading the glorification of AI, embracing religious metaphor to describe the power of the technology and themselves as nearly “gods creating new things,” said Epstein. He cited examples of the “fervor in the AI industry”: Twitter co-founder Jack Dorsey’s social media post, in which he said he believed in new owner Elon Musk’s “mission to extend the light of consciousness.” OpenAI CEO Sam Altman described his firm’s product as “magic intelligence in the sky.” Former Google and Uber executive Alex Levandowski created “The Way of the Future” church to worship artificial intelligence.
If AI becomes more powerful and more human-like, it won’t change that fact or even get close to providing an answer to that human quandary in the way that religion has done since the beginning of time.Matthew Ichihashi Potts
“There’s a danger in projecting divine goodness, or some transcendent intentions onto what is ultimately an extraordinarily large economic force that wants to become ever larger and evermore influential,” said Epstein. “It wants to sell more products; it wants to dominate more markets; and there aren’t necessarily benign intentions behind that.”
For religious leaders, AI’s ascent raises fundamental questions about the essence of humanity. The Rev. Matthew Ichihashi Potts, M.Div. ’08, Ph.D. ’13, Pusey Minister in the Memorial Church and Plummer Professor of Christian Morals, said the question of what it means to be human has implications for ethics, moral values, and religion.
Unlike machines, human beings must reckon with suffering, illness, and death throughout their lives, Potts said. And even if AI becomes more powerful and more human-like, it won’t change that fact or even get close to providing an answer to that human quandary in the way that religion has done since the beginning of time, he added.
“Whatever AI can tell us about how to reckon with the material reality that we’re subject to illness and death, us, humans, are still going to reckon with that,” said Potts. “That is when religion becomes significant.”
“To me, religion is primarily about coming into some meaningful relationship to one’s own vulnerability, to the fact that we’re human, that we suffer and that we’re finite,” said Potts. “That is what religion is for — to help us come into some meaningful relationship and also, hopefully some peace, with the facts of our finitude and vulnerability.”
Faith leaders and religion scholars advise AI users to keep a healthy dose of skepticism about the technology’s powers. Large language models allow AI to generate intelligent responses, but religion scholar Charles M. Stang wonders whether the technology is capable of actual intelligence as we understand it.
“AI requires building off a compendium of written knowledge, at least for most of the generative AI stuff, so there is already a bias towards religious faiths that have a large compendium of written knowledge.”Jenn Louie
“I still remain skeptical that AI is quote-unquote intelligent,” said Stang, Professor of Early Christian Thought at Harvard Divinity School. “To call our computers intelligent because they mimic a kind of intelligence we think we possess, possess exclusively, and regard as the pinnacle of intelligence strikes me as a form of technological narcissism.”
“In my experience, the tech world is prone to megalomaniac visions of the future and their role in ushering in the new age,” Stang said. “I don’t take that entirely seriously, but who knows? Maybe they will be the last ones laughing, or weeping.”
For Jenn Louie, M.Div. ’23, founder of the Moral Innovation Lab and a researcher at the Berkman Klein Center for Internet & Society, AI also poses ethical problems in the way it could disseminate information with a pro-Western bias and devoid of nuances.
“AI requires building off a compendium of written knowledge, at least for most of the generative AI stuff, so there is already a bias towards religious faiths that have a large compendium of written knowledge,” said Louie, who worked nearly two decades in the tech industry.
“If a compendium of knowledge is not written in English, which is where we’re advancing most of AI, will it always then bias us toward having morals and values that look and are very Judeo-Christian in nature? What are we doing about the practices that aren’t written in English?”
For all its potential evils, most faith leaders support the ethical use of AI. In 2020, the Vatican announced the Rome Call for AI Ethics calling for an ethical approach to AI. But they also say technology should not be treated as a source of religious inspiration or spiritual wisdom.
Epstein, the Humanist chaplain, said enlightenment can’t be found in algorithms, and technology can’t feel human emotions, understand pain or the importance of human connection. He urged people to be cautious when using AI without “trying to eliminate it and go back to the days of no tech.”
“We can use AI as a tool with certain strengths,” said Epstein, “but we must also look for ways that we can connect with one another on a purely human level as AI becomes more powerful and more dominant in our society. We can’t allow basic human connection to become a luxury item. Interfacing with tech is no excuse or no substitute for connecting deeply and meaningfully with one another.”
Hanching Chung 分享了 1 則動態回顧。
Simon 的大部分作品由 MIT Press 出版，包括其經濟學論文集多本者。這兩本認知心理學的論文集則由耶魯大學出版社出版 (1979/1989)，共千來頁。 謹以此小則筆記，紀念他並謝謝他的教誨。 (我們的二三十封通信，藏 Carnegie Mellon University圖書館數位檔。)
Herbert Alexander Simon (June 15, 1916 – February 9, 2001) was an American economist, political scientist and cognitive psychologist, whose primary research interest was decision-making within organizations and is best known for the theories of "bounded rationality" and "satisficing".[5] He received the Nobel Prize in Economics ……
我對於將論文集成書的朋友，總是推薦用Herbert Simon的作法：將書分成幾組，各組之前再補序言，使其成系統。
The Harvard Gazette
What we lose when we worship AI
Religion scholars and humanists have a warning about machines that know every detail of love and suffering except how they feel.
沒有留言:
張貼留言