Google announced today that Search Generative Experience (SGE) is now open for teens aged 13-17 in the US. Eligible users can sign up for Search Labs to try out SGE via Chrome desktop or Google app on their phone.
For the uninitiated, SGE is an experimental feature that displays AI-generated content on top of Google Search results. Announced during the Google I/O conference earlier this year, the generative AI feature is already available for users aged 18 and above. SGE was initially available in the US and made its way to other countries in the subsequent months.
In its blog post, Google stated:
Today, we’re opening up access to SGE in Search Labs to more people — specifically, teens (13-17) in the U.S. — so they too can benefit from the helpful capabilities generative AI has to offer. And as we bring SGE to more people, we’re continuing to deepen our investment in quality and safety, rolling out important updates and tools that improve the experience and help people find useful, reliable information.
One difference between conventional Google Search results and SGE is that users can throw follow-up questions at the AI to dig deeper. For that, users can click on the "Ask a follow up" option at the bottom of the AI snapshot to enter Converse mode and ask further questions related to their query.
Recently, Google has started to mention the website links that back up the information provided by its AI. Also, users can click on the Listen button to make Google read the AI-generated content aloud and hover over certain words to see their definitions.
The search giant said it "built additional safeguards into the experience" based on inputs from experts in teen development. For instance, SGE is designed to take extra precautions when dealing with outputs related to illegal or age-gated substances or bullying. With that said, Google cautions that SGE is still an experimental AI feature and things might go south sometimes.
Google is also rolling out 'About this result' for its search generative experience. Similar to how it works for regular search results, the features will give users more context on how the SGE generated a particular response. The company has plans to add "About this result" for individual links that show up in SGE results.
One area where we’re making targeted improvements is when a query includes a false or offensive premise – which can result in an AI-powered response that unfortunately appears to validate said premise
It's rolling out an update to help its AI model better detect such queries and respond with more accurate responses. In the future, the company wants large language models to "critique their own first draft responses on sensitive topics" and re-write their responses based on quality and safety principles.