Public librarians in Maine and Philadelphia are running workshops called Avoiding AI, and the sessions are drawing roughly six times the crowd of a standard tech-literacy class. Hannah Cyrus at Bangor Public Library capped registration at 30, opened a waitlist, and added a Zoom livestream — about 70 people attended each of her first two workshops. Her regular Intro to Computers classes normally get around 12.
Charlie Bailey, a librarian in South Philadelphia, followed Cyrus's model after she published a journal article on the format. His branch's Instagram post about the event pulled more than 2,000 likes and 220 shares, against a baseline of a few dozen. He scheduled a second session because the first hit its cap.
“I was inspired by the feeling of people's frustration with AI tools being kind of forced onto them, and feeling like AI tools we didn't ask for are suddenly everywhere in our lives.”— Charlie Bailey, Librarian, South Philadelphia
The workshop content is unglamorous and practical. Bailey opens with a plain-English explanation of how consumer AI features work, then projects step-by-step instructions for turning off Apple Intelligence, Gemini, and adjacent features across the major platforms. Attendees pull out their own phones and follow along. One attendee volunteered that appending "&udm=14" to a Google Search URL hides the AI Overview results — Bailey wrote the string on the whiteboard next to the teen Wi-Fi credentials.
Key facts
- 01Hannah Cyrus capped Avoiding AI registration at 30 in Bangor and opened a waitlist; roughly 70 people attended each of her first two sessions including Zoom.
- 02Cyrus' usual Intro to Computers classes draw about 12 attendees — the AI workshop pulled roughly 6x that.
- 03A Philadelphia library Instagram post for the workshop hit over 2,000 likes and 220 shares, versus a few dozen likes on typical posts.
- 04Bailey's Philadelphia workshop walks attendees through disabling Apple Intelligence and Gemini step-by-step on-device.
- 05Attendees shared workarounds including appending '&udm=14' to a Google Search URL to hide AI-generated results.
The demand pattern is what makes this a story. Cyrus said the requests started with patrons asking how to stop AI features from summarizing one-sentence emails or drafting replies they didn't ask for. Dozens of librarians from around the world emailed her for her slide deck after the journal article ran — a first, she said, for any program she's built.
Bailey framed the workshop as digital literacy rather than activism. As an information professional, he said, it feels good to see how many people share the skepticism, and the workshop is about giving patrons the mechanical knowledge to opt out where the product design makes opting out non-obvious. That includes navigating settings menus on iOS and Android that have been reorganized to make AI features the default surface.
Attendees are not uniformly anti-AI. Gabrielle, a workshop attendee in Philadelphia, said she keeps getting AI "shoved down my throat" at work and thinks about the environmental cost each time, but added that she is "not against AI in terms of medical breakthroughs." Another attendee, Johnny, raised the data-center siting question — the risk that a residential home purchase today could sit next to a hyperscale facility in two years. The concerns are specific and grounded, not abstract.
Cyrus was careful to distinguish rejection of AI from rejection of technology. She noted that optical character recognition — a machine-learning application — is genuinely useful for scanning old library documents. The pushback, in her framing, is about consent and defaults: whether users get to decide when AI mediates their communication, their search results, and their device behavior.
The broader signal is that the forced-adoption model — shipping AI features on by default across operating systems, search, email, and productivity apps — has produced a consumer counter-reaction organized enough to fill rooms at public libraries. That's not the same as a mass exodus from AI products, and the numbers here are small in absolute terms. But 70 people at a Bangor library workshop is a leading indicator that the default-on strategy has costs.
“I think the forced adoption of AI on people's devices might be the straw that's breaking the camel's back in some ways.”— Hannah Cyrus, Librarian, Bangor Public Library
The counterweight: none of this has shown up yet in usage numbers at OpenAI, Google, or Apple. Consumer complaint intensity and consumer switching behavior are different variables, and platform-level AI features are notoriously sticky once users stop noticing them. Cyrus and Bailey are also, by definition, self-selecting the frustrated: people who register for an Avoiding AI workshop are not a random sample of the installed base.
For the AI platform companies, the librarian workshops are worth watching less as a threat and more as a preference-revelation exercise. Users are telling public institutions, in specific detail, which AI features they want disabled — email summarization on short emails, unsolicited draft-writing, AI Overviews in search — and which they'd keep. That's a product roadmap hiding inside a complaint. The vendors that ship granular, easy-to-find opt-outs will hold users the default-on vendors are quietly losing.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
