All news
aiproductregulation

Librarians' 'Avoiding AI' Workshops Go Viral: What It Means

28 Jul 2026

A grassroots pushback on default AI features

A librarian in South Philadelphia, Charlie Bailey, recently hosted an hour-long workshop with an unusual title: "Avoiding AI." The session taught roughly 20 attendees how to disable Apple Intelligence and Gemini on their devices. It wasn't a one-off — demand was strong enough that Bailey scheduled a second session after the first filled up with registrations.

Bailey isn't the originator of the concept. That credit goes to Hannah Cyrus, a librarian in Maine, who developed the original "Avoiding AI" workshop and published a journal article about it. That article set off an unexpected chain reaction: dozens of librarians worldwide reached out to Cyrus after reading it. Her own first two workshops — capped at 30 in-person registrants — ended up drawing about 70 attendees each once livestream viewers were counted. For comparison, Cyrus's typical Intro to Computers classes draw about a dozen people. She described the level of interest as unprecedented relative to her other library programming.

The Philadelphia library's Instagram post about Bailey's event amplified the trend further, gathering over 2,000 likes and 220 shares.

Why this is happening

Bailey framed the workshop not as anti-technology, but as an exercise in digital literacy — helping people reclaim autonomy over whether and how AI tools are used on their own devices. The pattern across both Cyrus's and Bailey's events suggests a segment of the public wants more control over AI defaults rather than blanket opposition to AI itself.

The report notes this could reflect rising skepticism toward how Big Tech has rolled out AI features — often enabled by default rather than opt-in — and growing public interest in opting out altogether.

What's still unclear

Several open questions remain. It isn't known how many total "Avoiding AI" workshops have taken place globally, nor is there demographic data on who's attending (age, tech background, etc.). The workshops appear to cover Apple Intelligence and Gemini specifically, but it's unclear whether other AI platforms are addressed. There's also no data yet on whether attendees actually followed through and disabled AI features long-term, or whether this is a durable shift in sentiment versus a localized, niche phenomenon.

Why founders should care

This trend is a small but potentially telling signal. The workshops' rapid, organic spread — from one librarian's journal article to dozens of libraries reaching out, to viral social engagement — may suggest that a meaningful slice of consumers want more control over AI defaults, which could point to unmet demand for opt-in, rather than opt-out, AI product design.

For founders, this could mean:

  • Product design risk: Companies embedding AI features by default may face reputational friction or user pushback if this sentiment is more widespread than currently understood. This is a possibility worth monitoring, not a confirmed trend.
  • Market opportunity: The high turnout and viral sharing (2,000+ likes, 220 shares on a single post) hints at real appetite for tools and content addressing AI fatigue, privacy concerns, or transparency — a potentially underserved niche for privacy-first or opt-out-friendly AI products.
  • Distribution channel: The grassroots, librarian-led nature of this trend suggests community institutions like libraries and schools could be viable, low-cost venues for founders to pilot consumer-facing AI literacy or privacy-focused products.
  • Education gap: Cyrus's description of "unprecedented" outreach volume may point to a broader, currently unmet need for accessible, non-corporate AI education — a space startups could plausibly fill.

That said, the underlying data — no demographic breakdowns, no measure of actual long-term behavior change, and no confirmation this extends beyond a handful of libraries — means it is far from certain this represents a broad market shift versus a localized curiosity. Founders should treat this as an early, low-confidence signal worth watching rather than a validated trend to build a roadmap around.

Sources