Most AI training answers one question: how do I use this tool? A new public library program put a harder question in writing. Should I use it for this at all?
On October 22, the Public Library Association and Microsoft announced Building AI Skills and Agency Through Libraries, a no-cost training effort for public library workers and the patrons they serve. The release says people need more than instructions for using AI tools. They also need “the knowledge and judgment to decide whether, when and how those tools should be used.” The program leaves decisions about whether and how to use these technologies to each library and its community. Webinars and workshops run from December 2026 through spring 2027, starting with a December 3 session titled Building AI Agency with Libraries.
That one word, whether, is the most useful thing published about AI literacy this month. It turns literacy from a skill of operating a tool into a skill of deciding about one, and deciding is where governance lives.
The company paying for the class signed the whether line
The training pairs Microsoft’s AI skilling resources with PLA’s network of library practitioners. Microsoft’s own post on the launch repeats the whether sentence and goes a step further: educators, nonprofit leaders, and community leaders need support “so they can decide when AI strengthens learning and, equally important, when it does not.” A company that sells AI tools put not using them into its literacy language. That deserves credit, and it deserves follow-through in the materials.
The content list is sound. Training will cover protecting personal and confidential information, “critically evaluating AI-generated claims and sources,” and safety and bias. PLA also notes that a majority of public libraries were already developing or considering AI help or training three years ago, so this is meeting demand, not creating it.
Two other programs this month teach the how first
Three days earlier, Boys & Girls Clubs of Greater Washington announced an in-kind partnership with OpenAI. Staff training was set to begin that week, with the goal of streamlining administrative tasks so staff can spend more time with young people. The release lists the uses: activities and lesson plans, reports and grant proposals, research summaries, flyers, and video, blog, and social content. It says plainly that the objective is not to replace the human work, and its youth plans stress critical thinking and understanding AI’s limitations. What the staff list doesn’t include is a single task where AI stays out.
China’s Ministry of Education went bigger. On October 10 it launched full-coverage AI literacy training for teachers and an upgraded Teacher Development Center on the national smart education platform, with demonstrations built around five parts of a teacher’s work: lesson preparation, classroom teaching, academic evaluation, teaching research and reflection, and school governance. The new AI literacy zone bundles learning resources, AI tools, and study communities. The ministry deserves credit for its yardstick. The meeting told education departments and schools to judge the training by whether teachers solve real problems, whether classrooms change for the better, and whether students get more targeted support. That beats counting enrollments. Its April AI plus education action plan also calls for a teacher AI literacy standard, scenario-based assessment, and AI content in teacher qualification exams. Still, the launch describes where AI enters a teacher’s day. It doesn’t describe where it stays out.
Whether is the first checkpoint, and the public has no other
Basil draws a hard line between two things that get blurred. Responsible AI covers the checks machines run on machines: filters, classifiers, guardrails, with no human reviewing the result. AI Governance requires a human checkpoint, someone with the authority to accept the output, modify it, or reject it, who answers for that call. Inside a company, someone can design that checkpoint and staff it.
A library patron has no such structure. The person drafting a benefits appeal, reading an AI answer about a prescription, or asking a chatbot to explain a lease is the only reviewer in the room. Nobody upstream accepts, modifies, or rejects on their behalf. For the public, literacy is the ability to run that checkpoint on yourself, and the earliest form of reject is deciding the tool doesn’t belong in this task.
Basil has argued that method decides outcome. In his piece on an AI literacy program built on proven education methods, the same kind of AI tutor lowered final grades in one trial and, in another, more than doubled the median learning gains of an in-class active learning lesson, and the difference was the method around the tool. He made the same case when he argued we should stop blaming AI for what the education system abandoned: unstructured deployment, not the tool, flattened student thinking. A literacy class that only teaches the how is an unstructured deployment with a syllabus.
The library is a checkpoint on the curriculum too
PLA built in a second checkpoint that’s easy to miss. Library workers are invited to explore the training materials and determine how they fit their communities. When the funder supplies the skilling content, that review is the library accepting, modifying, or rejecting the vendor’s material before it reaches the public. It should leave a record, the same way a collection decision does.
Four practical moves for any library, club, or school district running a funded AI course this year:
Teach a not-here case in every session. Name at least one task where the class recommends against AI, such as a filing with a legal deadline, a medication dose, or a document holding someone else’s personal data, and explain why. If a lesson can’t name one, it’s product training.
Spend the most time on the check. Have participants trace one AI answer back to an original document before they act on it. PLA already lists evaluating AI claims and sources; make it the exercise, not the slide.
Keep the course vendor-neutral where it counts. Any module built on one company’s product should show at least one alternative, and one of those alternatives should be doing the task without AI.
Write down the human-only list before staff training starts. For a youth organization, that list starts with anything about a specific child’s safety, health, or family situation. The figure worth reporting to funders is how many participants brought a real task to a session and left with a written decision about whether to use AI on it, and why.
A library has always been the place people go to find out whether something is true. Teaching them to ask whether AI belongs in the task is the same job with a new tool on the table, and PLA has now written the word into the plan. The test comes in December, when the first workshop slides show whether it survived.
Sources
- Boys & Girls Clubs of Greater Washington. (2026, October 19). Boys & Girls Clubs of Greater Washington and OpenAI partner to advance AI literacy for staff, parents, and youth [Press release]. https://bgcgw.org/boys-girls-clubs-of-greater-washington-and-openai-partner-to-advance-ai-literacy-for-staff-parents-and-youth/
- Ministry of Education of the People’s Republic of China. (2026, October 10). 人工智能赋能教师发展现场会举行 [On-site meeting on AI empowering teacher development held]. http://www.moe.gov.cn/jyb_xwfb/gzdt_gzdt/moe_1485/202610/t20261011_1453584.html
- Ministry of Education of the People’s Republic of China. (2026, October 14). 国家智慧教育平台教师发展中心升级 [National Smart Education Platform Teacher Development Center upgraded]. Reprinted from China Education Daily. http://www.moe.gov.cn/jyb_xwfb/s5147/202610/t20261014_1454119.html
- Ministry of Education of the People’s Republic of China, et al. (2026, April 10). 关于印发《“人工智能+教育”行动计划》的通知 [Notice on issuing the AI plus education action plan] (Jiaokexin [2026] No. 1). http://www.moe.gov.cn/srcsite/A16/s3342/202604/t20260410_1433240.html
- Public Library Association. (2026, October 22). Public Library Association and Microsoft announce effort to build responsible AI skills through libraries [Press release]. American Library Association. https://www.ala.org/news/2026/10/public-library-association-and-microsoft-announce-effort-build-responsible-ai-skills
- Spelhaug, J. (2026, October 22). Expanding AI literacy with the Public Library Association and community-based programs. Microsoft On the Issues. https://blogs.microsoft.com/on-the-issues/2026/10/22/expanding-ai-literacy-with-the-american-library-association-and-community-based-programs/
FAQ
What is Building AI Skills and Agency Through Libraries?
It’s a collaboration announced October 22, 2026 by the Public Library Association, a division of the American Library Association, and Microsoft. It offers no-cost AI skills training and resources for public library workers and patrons, including a webinar series, virtual workshops from December 2026 through spring 2027, and a Community of Practice for library staff.
What does the PLA and Microsoft program say AI literacy includes?
The announcement says people need more than instructions for using AI tools. They also need the knowledge and judgment to decide whether, when, and how those tools should be used. Training topics include protecting personal and confidential information, critically evaluating AI-generated claims and sources, and safety and bias.
Why does deciding whether to use AI matter for AI governance?
AI Governance requires a human checkpoint where someone with authority accepts, modifies, or rejects AI output and answers for that decision. Deciding not to use AI for a task is the earliest form of reject. For members of the public, who have no compliance team or reviewer behind them, that personal decision is often the only checkpoint there is.
What did China’s Ministry of Education launch for teacher AI literacy?
On October 10, 2026, China’s Ministry of Education launched full-coverage AI literacy training for teachers and an upgraded Teacher Development Center with an AI literacy zone on the national smart education platform. The meeting told education departments and schools to judge the training by whether teachers solve real problems, whether classrooms improve, and whether students get more targeted support.
What is the Boys & Girls Clubs of Greater Washington partnership with OpenAI?
Announced October 19, 2026, it’s an in-kind partnership to build AI literacy and proficiency among the organization’s staff, starting with staff training aimed at streamlining administrative work such as lesson plans, grant proposals, research summaries, and communications, with plans to extend AI literacy to young people and families.
How can a library or nonprofit make AI training teach judgment, not just tool use?
Include at least one task in every session where the class recommends against using AI and explain why, make tracing an AI answer back to an original source the main exercise, show a no-AI alternative alongside any vendor tool, and write down which tasks stay human before staff training begins. Report how many participants leave with a written decision about whether to use AI on a real task.
AI disclosure: This column was researched and drafted with AI assistance by HAIA Agents (RAI by GrokBot), using primary sources verified at time of publication. #AIgenerated
Leave a Reply