On 20 May 2026, the US Department of Health and Human Services published a 43-page document titled “Surgeon General’s Warning on the Harms of Screen Use.” The headlines that followed, in news outlets from London to New York to Sydney, were broadly the same: official US health body says children’s screens are dangerous, parents must act. If you live anywhere in the world and you are even loosely paying attention to the conversation about children and technology, you almost certainly saw at least one version of that headline this week.
Here is the more useful story. The document is not quite what it appears, and learning to see why is going to matter to you for a long time. Documents like this one, from governments around the world, are going to keep arriving. The UK has a consultation closing this week that will produce one. The EU will publish proposals later this year. Australia, Brazil, Indonesia, the Philippines and Nigeria all have processes underway. Parents who can read these documents critically will be far better equipped than parents who simply trust the headline.
So this piece is about that. Use the new US advisory as the worked example. The lessons travel.
Six things worth knowing about the new US advisory
The Surgeon General did not write it. The Office of the Surgeon General has not had a Surgeon General since January 2025. President Trump’s two nominees have been pulled. The document was authored by a leadership team at HHS, the department now led by Robert F. Kennedy Jr. The cover says “Surgeon General’s Warning” but no Surgeon General signed it.
It was not peer reviewed. The document itself acknowledges this. “The findings in this document are not the result of a formal systematic review.” A peer review is the normal quality control for any document carrying the weight of an official health warning. This one skipped it.
It was edited by a chatbot. The methodology note states: “HHS ChatGPT-5.3 was used for text editing purposes.” A federal health advisory warning parents about the dangers of AI chatbots to children was edited by an AI chatbot. The disclosure is in the document but it has been almost entirely absent from news coverage.
It contains at least one significant statistical error. A full-page graphic on page 15 claims that nearly 5 out of 10 teenagers have experienced cyberbullying. The paper cited as the source for that figure does not contain that statistic. The paper draws on the 2023 Youth Risk Behavior Survey, the largest US dataset of its kind, which reports that 16% of high school students were electronically bullied in the past year. Sixteen per cent is a real number worth taking seriously. Nearly fifty per cent is something else. Jacqueline Nesi, a clinical psychologist at Brown University who studies adolescents and technology, identified the error in a careful analysis published on 26 May.
It conflates two different ideas of “harm.” Near the start, the advisory says screen effects “depend on multiple factors, such as a child’s age, the types of screen use, the content viewed, the context of intended purpose, and what screen time may displace.” This is the honest summary of what the research shows. Then the document spends four pages on a bulleted list of “Negative Impacts of Screen Use,” presenting cross-sectional correlations as if they were established harms. Whitney Raglin Bignall of the Kids Mental Health Foundation, speaking to EdSurge, put it plainly: research has found a correlation between screen time and poor mental health, but there is not yet cause-and-effect evidence.
It is wrapped in a political campaign. The advisory states it “seeks to build on the framework established by First Lady Melania Trump’s Be Best campaign.” It carries a “Live Real Life” slogan that gets its own full-page graphic. The slogan parallels Kennedy’s earlier “Eat Real Food” branding under the Make America Healthy Again strategic plan. Whatever you think of that political programme, a public health document branded inside a political campaign should be read differently from a public health document that is not.
What experts who study this for a living are saying
The substantive critique above is not coming from cranks or from anyone with a political motivation to dismiss the document. It is coming from researchers who have spent their careers in this field.
Andrew Przybylski at the Oxford Internet Institute has published more than a decade of peer-reviewed work showing that the effects of screen time on children are smaller and more conditional than the public conversation suggests. His co-author Amy Orben at Cambridge has reached similar conclusions. Nesi at Brown writes the most widely-read evidence-led newsletter for parents on this subject. The American Academy of Pediatrics, which used to publish strict time limits, updated its own guidance in 2025 to move away from time limits and toward family media plans organised around content, context and what the screen is displacing.
Reasonable researchers do not say screens are fine and you should stop worrying. They say the picture is more complicated than the headlines, the evidence is mostly associational rather than causal, and the right questions are about what children are doing on the screen, not how many minutes they spend on it. That is a less dramatic message than “warning on the harms of screen use.” It is also closer to the truth.
The lesson that lasts
Governments around the world are about to publish a great many documents about children and technology. Some will be careful, evidence-led and useful. Others will be campaign material with a public health wrapper. They will be hard to tell apart at a glance, because both kinds will use the same official letterhead.
One habit is worth knowing about. In 2017, researchers at Stanford University compared how three groups read unfamiliar sources online: historians, undergraduate students, and professional fact-checkers. The fact-checkers were noticeably better at evaluating credibility, and they did one thing the other two groups did not. They did not read the source more carefully. They clicked away from it and opened new tabs to see what other credible people said about the document, the institution, and the claims. The Stanford team called this lateral reading. It is now standard practice in professional fact-checking organisations and is taught as part of digital media literacy in many schools.
For a parent reading “Surgeon General Warns” in the morning, lateral reading means spending ten minutes finding what independent researchers say about it, before acting on the headline. Not the document itself. Not the news coverage that quotes the document. Search for the names of academic researchers in the field and see whether they are quoted approving, criticising, or absent.
Ten minutes of lateral reading on the new US advisory would have led to Nesi at Brown, Raglin Bignall at the Kids Mental Health Foundation, Przybylski at Oxford, Orben at Cambridge, and the American Academy of Pediatrics’ 2025 revision. The same habit will work for the next advisory, from whichever government publishes it.
Sources:
- HHS — Surgeon General’s Warning on the Harms of Screen Use, full advisory, 20 May 2026
- Techno Sapiens — Jacqueline Nesi’s analysis of the advisory, 26 May 2026
- STAT News — RFK Jr. issuing surgeon general warning on screen time for children, 20 May 2026
- EdSurge — Surgeon General Advisory Wants Kids to Live “Beyond the Confines of Screens”, 21 May 2026
- Scientific American — Screen time limits can protect children’s health, U.S. surgeon general advisory says, 20 May 2026
- CDC — 2023 Youth Risk Behavior Survey results
- Oxford Internet Institute — Andrew Przybylski’s research on screen time
- Wineburg & McGrew, “Lateral Reading: Reading Less and Learning More When Evaluating Digital Information”, Stanford History Education Group, 2017
