Skip to content
IFAI

AI tools

Google opens generative AI search to teens with extra guardrails

Published Reach 33

Google just expanded its generative AI search experience (SGE) to U.S. teens aged 13 to 17 through Search Labs. The rollout includes stronger protections against bullying and illegal substances, alongside an "About this result" feature to explain AI answers. Google is also training the model to stop validating queries with false premises.

Google opens generative AI search to teens with extra guardrails
Image: techcrunch.com

While adults are still debating AI safety, Google is handing its generative search engine directly to high schoolers.

On Thursday, Google announced that it is opening its Search Generative Experience, known as SGE, to teenagers aged 13 to 17 in the United States. Eligible teens signed into a Google Account can now opt in through Search Labs via the Google app or desktop Chrome.

Giving teens conversational search with training wheels

Google says the push down into younger demographics came naturally from user data. Since rolling out SGE, the company found that the feature scored its highest satisfaction ratings among 18-to-24-year-olds, who strongly prefer conversational querying over traditional keyword typing.

To safely expand that experience to minors, Google consulted teen development experts to build specific guardrails. Product Management Senior Director Hema Budaraju noted that the system introduces stricter filters against harmful outputs, with particular focus on bullying, illegal goods, and age-restricted substances.

Alongside the demographic rollout, Google is tackling a persistent model flaw: hallucinating confirmations. When a user asks a question loaded with a false or offensive premise, AI models often play along and validate the false claim, even if the underlying search index contains accurate data.

Google is now training its models to spot these premise-loaded queries and refuse to play along. It is also testing an internal loop where the model writes a first draft on sensitive subjects, critiques its own output, and rewrites it to meet safety standards.

To increase transparency, Google is bringing its standard "About this result" panel into SGE. This lets users inspect how an AI overview was put together, with plans to expand the panel to the individual source links cited inside the answer.

Why it matters

For younger students, conversational AI search can turn vague homework questions into guided research without forcing them to click through dozens of blue links. That said, it also introduces a real risk of over-reliance on synthesized answers during a critical stage of developing research skills.

For Google, this is an existential race for the next generation of web searchers. Younger audiences increasingly turn to TikTok and Instagram to find recommendations and answers. If Google does not hook teenagers onto conversational search inside its own ecosystem now, it risks losing the next wave of internet traffic entirely.

We think focusing heavily on premise-loaded queries is smart, as that is where naive or exploratory users get misled the most. However, teenagers are notoriously effective stress-testers. No matter how many critique layers Google builds into the model, keeping guardrails intact against playful jailbreaks will be an ongoing headache.

Sources (1)

  1. [1]
    techcrunch.com(opens in a new tab)Originaltechcrunch.com

Topic trackerChatGPT's threat to Google Search

Every story on this topic, newest first.

  1. Google brings AI search to 120+ new countries
  2. Google Search can now generate images and draft emails
  3. Google opens generative AI search to teens with extra guardrailsThis story
  4. Google brings its AI search to India and Japan
  5. Google turns SGE into a reading buddy for long web pages

Best of AI, daily