Norwegian Study Finds 9-to-18-Year-Olds Are Abandoning Google Search for AI Chatbots
Researchers in Norway found that children and teenagers increasingly treat conversational AI as their default answer source, and no one has measured what that costs their source-evaluation skills.
A study from Norway has found that users between the ages of 9 and 18 increasingly skip Google and other conventional search engines, turning instead to AI chatbots when they need to look something up. The researchers say the consequences of that swap for how young people learn remain largely unmeasured. The findings were reported by Norwegian SciTech News.
The distinction between the two tools is the heart of the concern. A search engine returns a ranked list of sources and leaves the judgment call to the user, who decides which link to trust and which to skip. A chatbot compresses that work into one synthesized answer, frequently with no visible path back to where the claims originated. Researchers quoted in the report single out that difference as the crux of the worry: the habit of weighing sources against each other, which schools have spent years trying to build, may be weakening before anyone has designed a way to detect the loss.
Norway offers an early vantage point. Minors there have among the highest rates of internet access and device ownership in Europe, so patterns that appear in the country tend to reach other wealthy markets within a few years. The age band in the study is also revealing. These are students old enough to conduct independent research for schoolwork yet young enough that their information habits are still forming, meaning a chatbot-first preference would follow them into university and then into the workforce.
The report stops short of claiming AI assistants give worse answers than search results, and that ambiguity is part of the difficulty. Chatbots can be quicker and more directly useful for a homework question, and in many cases they surface accurate information. But they can also assert falsehoods in the same confident tone they use for verified facts, and a user who never sees a source list has fewer cues with which to catch an error.
The authors describe the long-term effects on critical thinking and media literacy as unknown rather than proven. That is an unusual position for research on a technology already this widely adopted by minors, and it leaves the study's central question open rather than answered.
What comes next will likely split along two tracks. Expect more studies aimed at measurable outcomes, such as test scores, source-evaluation skills and time spent on research tasks, rather than self-reported habits. Expect schools and parents to begin writing rules for AI use that currently barely exist. The harder question is whether the comparison skills that search engines incidentally taught can be rebuilt inside a tool that answers first and cites later, or whether a whole cohort grows up without them. Norway's data will not settle that on its own, but it gives the rest of the world an early look at an experiment already in progress.