Across the country, libraries are quietly becoming the front line of a new kind of public conversation about artificial intelligence—one that isn’t centered on building models or chasing the latest app, but on resisting influence. In a pattern that’s showing up in multiple communities, librarian-led teams are running “Avoiding AI” workshops that draw unusually large crowds from people who say they’re tired of Big Tech shaping what they see, what they buy, and even how they communicate.
The demand is notable not just because it’s high, but because it’s broad. These sessions aren’t limited to the usual suspects—early adopters, policy wonks, or people already steeped in AI discourse. Instead, they’re pulling in everyday users: parents trying to manage what their kids encounter online, job seekers concerned about algorithmic screening and recommendation systems, retirees who feel overwhelmed by constant prompts and “helpful” automation, and community members who simply want more control over their digital lives. The common thread is frustration with the sense that AI is everywhere, often invisibly, and that opting out is harder than it should be.
Libraries, of course, have always been places where people learn how to navigate information systems. But the shift here is subtle and important. The workshops treat AI not as a distant technology trend, but as a practical layer of everyday life—embedded in search results, feeds, recommendation engines, customer service chatbots, voice assistants, ad targeting, and “smart” features that appear in apps without much explanation. The goal isn’t to demonize technology or pretend it can be fully avoided. It’s to help participants understand where AI-driven behavior shows up, what it does, and how to reduce its impact when they don’t want it.
What makes these workshops go viral is the promise of something rare in the current AI moment: agency.
A workshop format built around real-world choices
While each library adapts the program to local needs, the structure tends to follow a consistent logic. Participants start by mapping the AI touchpoints in their own routines—where they interact with platforms, what kinds of content they consume, and which tools they rely on for communication, shopping, or learning. From there, the session moves into practical strategies: how to adjust settings, how to limit data sharing, how to recognize when an interaction is being mediated by automated systems, and how to choose alternatives that keep humans in the loop.
This is where the librarian-led approach stands out. Librarians are trained to teach information literacy: how to evaluate sources, understand bias, verify claims, and find reliable information. In these workshops, that skill set is applied to a new domain—digital systems that increasingly behave like editors, curators, and persuaders. The emphasis is less on “AI is good” or “AI is bad,” and more on “AI is influencing you—here’s how to notice it and respond.”
In many sessions, the first breakthrough comes from reframing. Participants often arrive thinking they’re trying to avoid “AI” as a single thing. But the workshops explain that most people aren’t interacting with one AI model; they’re interacting with a stack of automated decision-making. That stack may include machine learning recommendations, ranking algorithms, personalization systems, spam filters, fraud detection, and generative tools that rewrite text or generate images. Even when a user isn’t directly using a chatbot, they may still be affected by AI-driven ranking and targeting.
Once that becomes clear, the conversation shifts from fear to tactics.
Understanding influence: the invisible layer
One of the most discussed themes is how AI affects online experiences. Participants learn that what appears in a feed or search results isn’t simply “what’s out there.” It’s what a system predicts they’ll engage with, based on past behavior and inferred preferences. That prediction can be helpful, but it can also narrow exposure, reinforce existing beliefs, and increase the likelihood of encountering content designed to keep attention rather than inform decisions.
Workshops often include guided demonstrations—showing how small changes in settings can alter what a platform serves. For example, participants may compare results across different browsers or profiles, observe how quickly recommendations adapt after a new search pattern, or see how “personalized” features can persist even when users believe they’ve turned off relevant options.
The point isn’t to convince participants that every platform is malicious. It’s to show that personalization is a form of influence, and influence can be managed. When people understand that the system is optimizing for engagement or conversion, they can better judge whether the output is aligned with their goals.
That’s also why these workshops tend to resonate with people who feel “fed up with Big Tech.” The frustration isn’t only about AI generating content. It’s about the broader experience of being tracked, nudged, and steered—often without meaningful consent or clear explanations.
Limiting exposure: practical steps that don’t require technical expertise
The workshops’ second major theme is strategies for limiting exposure to AI-driven tools. This is where the sessions become especially useful for non-technical attendees. Rather than asking people to install complex software or learn programming, the workshops focus on settings, permissions, and behavioral habits that reduce data collection and personalization.
Common topics include:
1) Reducing data shared with apps and platforms
Participants review privacy controls such as ad personalization settings, location permissions, and account-level data sharing options. They also discuss how “default” settings often favor data collection, and how to check what’s enabled across devices.
2) Managing cookies and identifiers
Many workshops cover the difference between clearing cookies and actually preventing tracking. Attendees learn that some identifiers persist through logins and device-level signals, so “I cleared my history” doesn’t always mean “I stopped being profiled.”
3) Using separate profiles or browsers
A recurring tactic is separating activities—using different browser profiles for different purposes, or keeping a “research” profile that isn’t tied to daily browsing. This helps reduce cross-contamination of interests that feed personalization systems.
4) Rethinking sign-ins
Workshops often encourage participants to consider whether they truly need to be logged in to every service. Being signed in can increase the system’s ability to connect behavior across contexts.
5) Adjusting notification and recommendation settings
People are frequently surprised by how much recommendations can be shaped by notification preferences and engagement patterns. Turning off certain types of prompts can reduce the frequency of algorithmic nudges.
6) Recognizing AI-mediated interactions
Participants learn to identify when they’re interacting with automated systems—customer service chatbots, “smart” search suggestions, or auto-generated responses. The goal is not to avoid automation entirely, but to know when a human isn’t behind the interaction and to adjust expectations accordingly.
These steps are framed as “control measures,” not as a moral stance. The workshops emphasize that avoiding AI completely may be unrealistic, but reducing its influence is achievable—and measurable.
Building media and tech literacy grounded in user control
Another reason these workshops spread is that they offer a coherent literacy framework. Instead of treating AI as a mysterious force, the sessions teach participants to think like evaluators: What data is being used? What incentives drive the system? What is the likely objective—informing, persuading, selling, retaining attention?
Media literacy has long focused on verifying claims and understanding framing. Tech literacy has focused on understanding how systems work. These workshops blend both, with a distinct emphasis on user control. Participants practice questions such as:
What am I being shown, and why might the system think I’ll like it?
What information is missing because the system is optimizing for engagement?
How would I verify this claim outside the platform’s recommendation engine?
If I change my behavior slightly, does the system change what it serves me?
This approach is particularly effective for people who feel overwhelmed. It gives them a mental model that doesn’t require deep technical knowledge. They don’t need to understand neural networks to understand that ranking systems can shape reality.
And importantly, the workshops treat literacy as a skill that can be practiced. People leave with checklists and habits, not just opinions.
A space to ask questions—and compare experiences
Libraries also provide something that online discourse often lacks: a room where people can ask questions without being mocked or dismissed. In many communities, AI conversations have become polarized. Some people treat AI avoidance as naïve; others treat it as a moral imperative. Workshop attendees often report that the library setting feels different—more grounded, more practical, and more respectful.
That social element matters. When participants compare experiences—“I turned off this setting and my feed changed,” or “I tried this and it didn’t work”—they build collective knowledge. Librarians facilitate these discussions, helping participants interpret what they’re seeing and avoid jumping to conclusions.
This is also where the “unprecedented demand” becomes understandable. People aren’t just looking for information; they’re looking for reassurance that their concerns are legitimate and that there are steps they can take.
The unique role of libraries in the AI era
The most striking aspect of this trend is the setting itself. Libraries are not typically associated with hands-on guidance about AI avoidance. Yet they’re uniquely positioned to deliver it.
First, libraries are trusted institutions. People may not trust a platform’s own privacy messaging, but they often trust a library’s educational framing. Second, libraries are built for instruction. They have staff who can teach, materials that can support learning, and spaces designed for community engagement. Third, libraries serve diverse populations, including people who may not have access to paid privacy tools or advanced technical support.
In other words, libraries are translating a complex, fast-moving topic into accessible public education.
There’s also a policy dimension. When communities ask for AI avoidance workshops, they’re implicitly raising questions about transparency, consent, and accountability. If people feel they need training to reduce AI influence, that suggests the default user experience is not designed for informed choice. Libraries can’t rewrite platform policies, but they can help citizens understand what’s happening and advocate for better standards.
A “avoidance” mindset that doesn’t ignore reality
It’s worth noting that these workshops generally