Digital Literacy: What Schools Mean and What Kids Need

The phrase turns up in school reports, curriculum letters and screen time policies, usually without a definition. It covers so much ground, from typing to privacy settings to spotting a fake video, that as written it tells you almost nothing about what your child can actually do.

So here it is broken into its parts, with an honest account of which ones matter most now. The short version is that children are generally good at using technology and much weaker at judging what it tells them, and the newest version of that problem, judging what an AI says, is the part least taught of all.

What does digital literacy actually mean?

The ability to find, judge, make and share information using digital tools. It is a thinking skill as much as a technical one, which is the part most definitions bury.

The clearest definition comes from the American Library Association, whose Digital Literacy Task Force defines it as "the ability to use information and communication technologies to find, evaluate, create, and communicate information, requiring both cognitive and technical skills."

Notice the four verbs, because they are the whole idea. Find, evaluate, create, communicate. A child who is fast on a tablet has the technical half. Whether they can evaluate what they found is the cognitive half, and it is the one that does not come from practice alone.

What do schools mean when they use the phrase?

Usually a bundle of separate skills taught under one heading. Many US schools organize this around the ISTE Standards for Students, which split it into several distinct strands rather than one subject.

The ISTE Standards for Students include a Digital Citizen strand, in which "students recognize the responsibilities and opportunities for positively contributing to their digital communities," covering digital footprint, online interactions, well-being and privacy. A separate Knowledge Constructor strand asks that "students critically curate a variety of resources using digital tools to construct knowledge," and its parts include evaluating and curating information.

That split explains why a school report can say your child is doing well at digital literacy while they still believe the first search result. The report may be describing safe and polite behavior online, which is real and worth having, while saying nothing at all about whether they can tell a reliable source from a convincing one.

It is worth asking which strand a comment refers to. "Good digital literacy" about online behavior and "good digital literacy" about evaluating sources are describing two different children.

Which parts matter most for a child now?

Judging information, ahead of everything else. Typing, file handling and knowing which button does what are real skills, but children pick most of them up simply by using devices, and they rarely hold a child back.

The judging part does not develop on its own, and the evidence on how well teenagers do it is sobering. It is also the part that has changed most in the last few years, because the thing children are now asking questions of is often an AI system rather than a search engine.

Privacy and online behavior sit close behind. They matter a great deal, and schools usually teach them reasonably well because they map onto safeguarding, which schools take seriously.

How good are students at judging what they find online?

Not good, even with a live internet connection in front of them. The largest study of the question found high school students struggled on every task it set.

Researchers at Stanford, in what they describe as "the largest investigation of its kind," assessed 3,446 high school students chosen to reflect the demographic profile of US high schoolers. Asked to investigate a site claiming to publish factual reports on climate science, "96% never learned about the organization's ties to the fossil fuel industry." Two-thirds "were unable to distinguish news stories from ads on a popular website's homepage." And "over half believed that an anonymously posted Facebook video, shot in Russia, provided 'strong evidence' of U.S. voter fraud."

The finding that matters most for a parent is what fooled them. Students "were often duped by weak signs of credibility: a website's 'look', its top-level domain, the content on its About page, and the sheer quantity of information it provided."

Those are exactly the signals children are usually taught to check. A professional-looking site, a .org address and a detailed About page all feel like evidence of reliability, and none of them are. That is the uncomfortable part: some of the standard advice is the problem.

One caveat on scope. This studied high school students in the US, not younger children. It is reasonable to assume younger children do no better, but the study does not measure them.

What do professional fact-checkers do differently?

They leave the page. Instead of studying a site to judge it, they open new tabs and find out who is behind it from somewhere else, which the researchers call lateral reading.

The same study describes the contrast directly. Students and academics tended to read vertically, starting at the top of a page and judging its prose, layout and domain. Fact checkers "took a radically different tack. Landing on an unfamiliar website, they clicked away from it and engaged in lateral reading." They opened new tabs to search for information about the organization or person behind it, and only returned to the original page afterward.

The researchers boil what fact checkers ask down to three questions, and they are the most practical thing on this page. Who is behind the information? What is the evidence? What do other sources say?

Those three work at any age and require no technical skill at all. A ten-year-old can ask who made a video. A fourteen-year-old can search the name of the organization behind a claim before believing it.

Why is judging AI answers the newest and least taught part?

Because an AI answer has no page to leave. Lateral reading works by stepping away from a source to find out who is behind it, and a chatbot answer arrives with no author, often no sources, and a confident tone regardless of whether it is right.

The companies building these tools say so themselves. OpenAI's own Terms of Use state that "Output may not always be accurate. You should not rely on Output from our Services as a sole source of truth or factual information." That is the maker of the tool telling users not to trust it on its own.

So the fact-checker's third question becomes the whole skill. What do other sources say? A child who checks an AI answer against a second, named source before using it is doing digital literacy properly. A child who pastes it straight into homework is not, however fluent they are with the tool.

Our guide to ChatGPT for kids covers the age rules and the checking habit in detail, and our parents' guide to AI covers the wider literacy side.

How do you teach this at home without a curriculum?

Ask the three questions out loud, about things your child actually sees. A video, a claim a friend repeated, an answer an AI gave. Who made it, what is the evidence, what does anyone else say.

The habit forms from repetition in real situations, not from a lesson. When your child tells you something surprising they read, ask where it came from, then look it up together in a new tab. That is lateral reading, and doing it with you twice a week teaches more than a worksheet.

It also helps to retire some of the old advice. Telling a child to trust .org sites or to check the About page teaches them to trust exactly the signals the Stanford study found were fooling students. Teach them to find out who is behind a site from somewhere other than the site itself.

Where does learning to code fit into digital literacy?

It covers the create half, and it changes how a child understands the rest. A child who has built even a simple program knows that software does what someone decided it should do, which is the beginning of asking who decided.

That matters most with AI. Understanding that a model predicts likely text rather than looking up facts explains why it can be confidently wrong, and our explainer on how AI actually works covers that without assuming any background.

It is also why our Senior Track teaches AI alongside Python rather than on its own. A teenager who can already write code can understand what a model is doing rather than just typing prompts, which is a much sturdier form of AI literacy than knowing which prompts work.

If you would like to see how that is taught, a free trial class is one live session with a mentor, no card details and no commitment.