Google let teens access generative AI in search. They were probably already using it anyway

Must read

Google had prohibited a key audience—teens—from its generative AI search product that it introduced in May. Now, it has opened the floodgates to those younger users. 

On Thursday, the company extended the feature to users aged 13 to 17 in the U.S., as long as they have a Google account and opt in. The technology, called Search Generative Experience (SGE), offers AI-generated answers to search prompts and lets users browse web pages, with AI pointing to the most relevant information. 

Teens had been using Google’s AI products and similar ones for some time, notoriously creating an epidemic of AI ghost writing school essays. Online age restrictions are largely futile since most sites, including Google, have users self-report their ages.

It’s unclear why Google waited to officially open the product for younger users, but the Thursday announcement also featured new safety updates that coincided with the roll-out for teens. AI chatbots have repeatedly spit out responses filled with racism and other problems, and subjecting teens to it in the first few months—before some kinks in the technology were fixed—could have been a PR nightmare. Bard, Google’s own chatbot, spread misinformation when tested by researchers earlier this year. It spouted homophobic content and Holocaust denial, the Daily Beast reported. SGE is a different AI product from Bard, but the controversy shows why companies may take a cautious approach when releasing it to a young audience. 

Gen Zers and millennials are “super-users” for generative AI, accounting for 65% of total users, Salesforce data shows. But Google was missing a big section of Gen Z by excluding teens from its technology. By opening it to even younger users, the company can reach a new audience, which could translate to future profits as it expands its AI ad business.

“They too can benefit from the helpful capabilities generative AI has to offer,” Hema Budaraju, Google’s senior director of product, said of teens in a company blog post

Children under the age of 13 who aren’t supposed to create their own Google accounts and access SGE can still easily bypass age limits by changing their birth date in their account settings. Google says it can ask users to verify their age by providing a government ID for restricted services, but when tested by Fortune, a new user who had initially reported their age as 17 was able to change it to 21 and begin using SGE immediately. 

Teens have access to an array of generative AI technologies, not just Google’s. With parental consent (in theory), they can get access to OpenAI’s ChatGPT chatbot. Microsoft lets teens search with Bing AI—Google’s competitor. Snap has My AI for those 13 or older, and Meta on Wednesday announced new chatbots on 13-and-up Instagram, WhatsApp and Facebook Messenger that replicate the personalities of celebrities like Kendall Jenner and Snoop Dogg. 

While teens will see ads in their SGE results, Google will not employ ad targeting on the group, a Google spokesperson told Fortune via email. This aligns with the policy Google implemented in 2021 to prohibit targeting based on age, gender, and interests for users under the age of 18 across all its products. While ad targeting for teens isn’t illegal in the U.S., it is in some other countries Google operates in, and the company decided to comply across its whole business for consistency, it said in a blog post.

The roll out comes with safety features intended to protect this more vulnerable population, Budaraju said in the post. SGE is designed to prevent inappropriate or harmful content from appearing, but for the teen experience, Google has implemented stronger safeguards against illegal material, bullying, and age-gated content, like alcohol products.

For all users, the company is adding an “About this result” feature for AI-generated responses, which will give context about how SGE produced its answer. Google is also working on how its technology responds to offensive queries from users, Budaraju said, which includes having the AI critique and rewrite its own draft response for sensitive topics.

Subscribe to Well Adjusted, our newsletter full of simple strategies to work smarter and live better, from the Fortune Well team. Sign up today.

More articles

Latest article