Librarians Hosting Viral 'Avoiding AI' Workshops

A Philadelphia library's Instagram post about a workshop on disabling AI features racked up more than 2,000 likes in July 2026 — against a typical baseline of "a few dozen" — and the shockwave it sent through the library world has turned a quiet digital literacy effort into a nationwide movement. Librarians from Maine to California, and now internationally, are running sold-out "Avoiding AI" sessions for people who are tired of having machine-learning features switched on without their knowledge or consent. The numbers are stark enough that the tech industry should be paying attention.

Background You Need

For the past three years, the major consumer technology platforms have been racing to embed AI assistants directly into operating systems, search engines, and productivity apps. Apple shipped Apple Intelligence as a default feature on iOS 18 and later devices. Google added AI Overviews to the top of nearly every search results page, replacing the link-list format that hundreds of millions of people had relied on for two decades. Microsoft wired Copilot into Windows at the system level. In each case, the rollout was opt-out rather than opt-in — meaning users who wanted the old behavior had to find and toggle an often-buried setting.

The backlash that followed was not initially loud, but it was persistent. Forum threads about disabling AI summaries accumulated hundreds of thousands of views. A tip about appending &udm=14 to any Google Search URL — which forces the engine to return a traditional link-based results page and hides the AI Overview panel — spread through social media and productivity communities as a kind of folk remedy. That URL trick, sourced from the community and passed hand-to-hand, eventually found its way into the library workshops as standard curriculum.

Into this environment stepped librarians, an institution with a century-long track record of teaching information literacy without a financial stake in any particular platform. The same credibility that made libraries the trusted venue for "how to spot a phishing email" classes in the 2000s is now being redirected at a different kind of unwanted intrusion: AI features that arrive pre-activated on devices people already own.

What's New

The proximate catalyst appears to be a journal article written by Hannah Cyrus, a librarian in Maine, which outlined a framework for running community workshops focused specifically on AI avoidance rather than AI adoption. The practical curriculum covers how to turn off Apple Intelligence on iPhones and Macs, how to disable Google's Gemini assistant on Android, and how to navigate the settings menus of other major platforms to restore something closer to the pre-AI experience. Multiple reports confirm that Cyrus's workshops scaled from intimate 12-person in-person classes to 70-person Zoom sessions with active waitlists — a growth curve that reflects real demand rather than promotional momentum.

The social media signal is equally striking. Where a typical library programming post might earn engagement in the double digits, posts about the "Avoiding AI" format are registering thousands of interactions. The Philadelphia library's 2,000-plus Instagram likes is the most concrete number in circulation, and multiple sources cite it as emblematic of a broader pattern rather than a one-off outlier. According to TechCrunch's reporting, librarians from outside the United States have since contacted Cyrus and others to ask how to replicate the model in their own communities — suggesting this is no longer purely an American phenomenon.

What makes the framing of these workshops interesting, and probably explains much of their appeal, is the explicit distinction attendees draw between AI they chose and AI that chose them. Workshop participants, according to the primary reporting, are not opposed to AI in oncology research or climate modeling. What they object to is a summarization widget condensing a one-sentence email into a one-sentence summary, or a search engine inserting a generated paragraph between them and the blue links they were looking for. Cyrus described forced adoption of AI on personal devices as potentially "the straw that's breaking the camel's back" — a formulation that positions this as a tipping point in a longer accumulation of grievances rather than a sudden ideological conversion.

Our read is that this distinction matters a great deal for how developers should interpret the signal. The people showing up to these workshops are not technophobes — they are users who had a functional relationship with their devices and felt that relationship was unilaterally altered. That is a product trust problem, not a technology acceptance problem, and it is a meaningfully different diagnosis.

The Pushback

The obvious counterargument is that most of these features can, in fact, be turned off — the platforms provide settings, even if they are not always easy to find. Critics of the "Avoiding AI" movement might argue that a workshop teaching people to navigate opt-out menus is solving a UX discoverability problem, not making a political statement, and that framing it as resistance overstates what is happening. Some in the tech industry will point out that AI features represent genuine utility for many users and that the opt-out architecture is a reasonable default when the majority of a user base benefits from a feature.

That argument is harder to sustain when a library workshop on the subject is generating thousands of social media engagements and growing internationally. Volume of interest is not the same as a policy verdict, but it is evidence that the discoverability argument misses something. If burying an opt-out setting were the only issue, the workshops would be attended by the chronically confused rather than by people who are, by all accounts, specifically motivated and showing up on waitlists. The frustration appears to run deeper than a settings menu — it is about the presumption embedded in an opt-out default: that a company's judgment about what belongs on your device supersedes your own.

Whether libraries can sustain this role long-term is also worth watching. Library systems are under significant funding pressure in many states, and programming budgets are not immune to cuts. An institution that has just discovered it can draw 70 people to a Zoom session about phone settings is well-positioned to expand digital literacy programming — but only if it has the staffing to do so. The viral moment is real; whether it translates into durable institutional capacity is a separate question.

Sources