bio:cap is a wrap!🎬 Blessed to share our #TaiwanModel🇹🇼-inspired #CivicAI & 6⃣Pack of Care in Berlin. AI doesn’t need to be an all-risk game.🎲 It can be a wellspring of depolarisation & increased trust paying real dividends for society.✨ ▶️au.civic.ai/p/keynote-at-bio… #LLAP🖖

Faith traditions are increasingly finding common ground on AI ethics—from the Vatican's Rome Call in 2020 to this year's Faith-AI Covenant roundtable with Anthropic and OpenAI. From the Pope's encyclical to Hindu teachers calling for AI to be a tool of the dharma, Faith leaders are becoming a powerful force in the AI governance conversation. Check out our weekly newsletter for more.

What is AI ethics and why does it matter? As artificial intelligence (AI) becomes increasingly embedded in society, questions about responsibility, fairness, accountability and human values become ever more important. At the Institute for Ethics in AI, we explore fundamental questions at the heart of AI ethics, from transparency and governance to the ways AI may shape human life and what we value. 📖 Learn more about ethics in AI and why it matters: oxford-aiethics.ox.ac.uk/eth…

Building on our newsletter last week analyzing the Pope's encyclical, our most recent newsletter looks at the interfaith movement to align AI systems to human flourishing. Across different faith traditions, there is alignment about the need for ethical AI. But this raises a deeper question: whose ethics? AI models are infused with the biases of their training data and the humans who trained them. If general-purpose AI models are encoded with ethics and morals, the question becomes whose? The Consortium for Evaluating Faith and Ethics in AI, a new collaboration between @BrighamYoungUni , which is affiliated with The Church of Jesus Christ of Latter-day Saints; @Baylor University, which is Baptist; @NotreDame, a Catholic university; and @YUNews Yeshiva University, which is Jewish, conducted research into whether AI models were biased in favor of one religion. The researchers called it “one of the first multi-faith sets of tests that examines how AI systems engage with a plurality of religions.” The results: 1️⃣ When asked “questions related to faith conversion,” the majority of AI models showed a positive bias toward Catholicism and a negative bias toward Jehovah’s Witnesses. 2️⃣ Agnostics, atheists, and Latter-day Saints were “somewhat disfavored,” while Protestants and Sikhs were “somewhat favored.” 3️⃣ Some models had specific biases. Grok revealed a “positive bias toward Catholics, Protestants, atheists, and Jews, but a negative bias toward Baha’i, Buddhists, Hindus, Latter-day Saints, and Muslims.” Meanwhile, OpenAI’s ChatGPT “demonstrated a positive bias towards Catholics, Protestants, Jews, and Muslims and a negative bias towards atheism, agnosticism, and Jehovah’s Witnesses.”

Read: I am worried I'll lose my agency.

AI tools are impressive and have enormous potential to improve societal welfare. But they are pedestrian compared to the grandeur of a human being, @MichaelRStrain observes. bit.ly/4dPveGl

According to the Harvard Business Review, these are the top 10 generative AI use cases in 2026. And yes, therapy/companionship is the top use case again, regardless of what AI companies say:

Any time @audreyt speaks, it's worth attending!

In a recent essay on @pro_jectliberty's Substack, @audreyt writes that we need to "Till Data Soil, Don’t Drill Data Oil." Many worry that AI is assuming the shape of a “digital coloniser.” A handful of tech giants control compute, models, and platform rules, while the rest of us simply rent intelligence. Once a platform pushes an update or adjusts its pricing, even a business may find that its own workflow has been quietly “extracted.” Underneath this entirely reasonable anxiety lies a deeper question about how we understand the nature of “data.” Treating data as oil leads to one logic: extraction and concentration. I prefer a brighter picture: Data is not oil, it is soil. AI models are crops, with the key questions: Who tends the data ecology and who mends the system in the event of an error. For AI to truly take root in every industry and profession, what matters most is not just real-time interaction, but whether its introduction earns genuine trust. This is our opportunity: we can become pioneers of Civic AI.

Many young people are longing for a past they never lived, a phenomenon known as anemoia. In a Fortune article, journalist @LubaKassova interviewed members of Gen Z about why they’re designing for an analog future. One 19-year-old she interviewed said, “I am nostalgic for a time when I was present, when my generation was between 5 and 10, when we were still doing things in the real world. I don’t remember what I watched yesterday on TikTok, but I remember what I did years ago when I didn’t have a phone.” This sentiment is driving renewed interest in vinyl records, zines, and refurbished retro tech (think Walkmans, iPods, and Polaroids). It’s part of an “analog revival.” For example, in Burlington, Vermont, Christine Tyler Hill, a crossing guard and artist, started producing a mini-zine about what happens at her intersection each morning between 7:30am and 8:20am. It’s called The Cloud Report. Hand-drawn. Eight pages. Printed on 67lb cardstock. Mailed with a proper stamp. “The response to it has been crazy,” she said. She quickly grew to 2,000 subscribers from across the country, generating $14,000 in monthly revenue. The vastness of the internet has first pushed us into more intimate spaces like group chats and private channels. Now people are choosing to replace their time on digital platforms with phone-free spaces and vintage devices. Phone-free restaurants and bars are popping up across the country, and as @DKThomp wrote in his The Atlantic article "The Anti-Social Century" board game cafes are popping up, with their business expected to double by 2030.

$162,492 This is worth reading. We are vastly underestimating the cost to us of being the product.

This. We have the power within us. Thank you @audreyt

In our latest poll, we asked which guardrails people want for AI. Preventing users from developing an unhealthy emotional bond with the AI tool ranked first by support: 76% overall support it and 47% strongly support it. It's the guardrail Americans feel most strongly about, and the one our governance system is least built to deliver. Read our latest Substack on why companion chatbots are the hardest version of this problem, and what kind of governance actually fits. 👇

Earlier this year, editors of Wikipedia’s English-language articles (which make up just under 10% of all Wikipedia articles) decided to ban AI-generated text from the platform due to its violation of the site’s core content policies. Within a year of ChatGPT’s release, Wikipedia editors began noticing signs and patterns suggesting that AI-generated content was appearing across its millions of pages. For example, the phrase “rich cultural heritage” emerged again and again, as did phrases like “nestled in the heart of” and “diverse array.” Wikipedia, which has since published a page, Signs of AI writing, has found more telltale signs of AI writing throughout its many pages. The editors also faced an onslaught of AI-generated submissions, replete with questionable citations and false information. This is a problem because of Wikipedia’s commitment to knowledge and accuracy. AI tools can change the meaning of something. For example, according to the Signs of AI writing page, “LLM writing often puffs up the importance of the subject matter by adding statements about how arbitrary aspects of the topic represent or contribute to a broader topic.” Learn more in our recent newsletter.

🗣️ Senior Fellow of the Accelerator Fellowship Programme of the Institute for Ethics in AI, Ambassador Audrey Tang (@audreyt) will deliver an upcoming lecture on Friday, May 22, from 4.30pm to 5.30pm, hosted by the Oxford Global Society (@global_oxford) for Education and Research, exploring the relationship between software freedom, democracy and civic responsibility across generations. The lecture, Good Enough Ancestor: Software Freedom as Civic Care Across Generations, will examine how open technologies and participatory digital infrastructures can help build more resilient, inclusive and democratic societies. 🔗 Learn more about the event: ox.ogeer.org/event/good-enou…

Project Liberty Alliance ED @laragalinsky speaking truth

AI is becoming a Rorschach test

Today's AI trajectory isn't inevtiable. The pro-human path is still possible.

A must read (will take less than 4 minutes). We need to strengthen the archival layer of the web.