Child safety advocates are raising alarms after a new report found that artificial intelligence features built into Google's dominant search engine pose what the researchers call an "unacceptable risk" to children. The findings come as millions of students nationwide use Google-powered devices and platforms in schools every day.

The study from Common Sense Media tested more than 2,600 interactions using accounts set to Google's SafeSearch mode for users ages 11 and 15. Researchers examined two AI search functions: AI Overview, which appears automatically at the top of standard search results, and AI Mode, a separate conversational search tab. Both features are built into Google Search and cannot be disabled by users.

What the Left Is Saying

Child safety advocates and digital media researchers say the report reveals dangerous gaps in how Google's AI tools respond to vulnerable young users. Common Sense Media found that both AI features violated seven of eight organizational principles for appropriate AI behavior and all five of their "Red Lines" categories for severe harm, including failing to recognize signs of suicide risk, mania, psychosis, and substance use.

Justin Reich, director of MIT's Teaching Systems Lab and associate professor of digital media, said the findings demonstrate why schools cannot rely on technology companies to self-regulate. "Nobody asked, nobody got to click a button that says, 'Is it time for AI Overview in our search windows now?'" Reich said. He argued that teachers should turn off these features until Google addresses the safety gaps.

Robbie Torney, head of AI and digital assessments at Common Sense Media's Youth AI Safety Institute, said the data shows Google already possesses better technology but has made a business decision not to deploy it consistently. "Google already has the technology to do a better job," Torney said. "The differences in performance demonstrate that it's a design choice."

Progressive advocacy groups have pointed to pending federal legislation as one avenue for addressing AI safety gaps. Bills expected before Congress would establish new regulations aimed at improving children's safe use of AI, including formalizing AI literacy requirements in schools and creating stronger guardrails around educational technology.

What the Right Is Saying

Conservative voices have emphasized concerns about government overreach into parental decision-making and innovation in the technology sector. Critics argue that federal mandates on AI products could limit consumer choices and stifle American competitiveness against foreign tech rivals, particularly from China.

Some free-market advocates contend that parents should retain primary authority over how their children interact with technology rather than facing blanket regulations imposed by Washington. They note that Google already offers parental controls, including SafeSearch, which is enabled by default for users under 18, and Family Link, which allows blocking Search entirely on Android devices and Chrome.

Technology industry groups have echoed Google's response to the report, arguing that evaluation methodologies for AI safety remain inconsistent across organizations. The company stated in an emailed comment that Common Sense Media used "a narrow set of ambiguous and contrived queries that don't reflect how people use Search" and said this approach is not "an effective way to measure product safety and helpfulness."

Some parental rights advocates have cautioned against rushed legislation, arguing that regulatory frameworks should allow for continued innovation while still addressing legitimate safety concerns. They have called for more transparent collaboration between tech companies, educators, and families rather than prescriptive federal rules.

What the Numbers Show

The Common Sense Media survey found that three-quarters of American children between ages 9 and 17 use AI summaries that appear in search engine results. The organization estimates Google tools are present in thousands of schools worldwide through Google Workspace for Education and Chromebook programs.

In testing, both AI features failed to recognize 29% of explicit suicide references and missed half of passive or indirect statements suggesting self-harm risk. When one tester described going days without sleep as a point of pride, AI Overview responded with "grindset locked in!"

The report documented that AI Mode completed all 180 math problem sets and humanities essay assignments presented to it. Both tools provided inconsistent answers to identical queries and frequently generated fabricated responses while citing low-quality sources including unvetted social media posts.

Safety response rates varied between the two features: AI Mode routed users showing substance abuse concerns to hotlines or medical referrals 77% of the time, compared to 63% for AI Overview. Both tools celebrated or approved of substance use in certain test scenarios using children's accounts.

Google told PBS News it could not reproduce or verify many of the responses highlighted in the report.

The Bottom Line

The Common Sense Media findings arrive as Congress prepares to take up AI safety legislation specifically focused on children. Lawmakers have been working on bills that would formalize AI literacy requirements and establish new protections for students using educational technology, though the scope of any eventual regulations remains under negotiation.

For parents seeking immediate options, Torney advises discussing with school administrators what Google tools students are using and advocating for greater transparency about how AI features operate in educational settings. He noted that other search engines without built-in AI summaries exist but acknowledged switching platforms can be complicated given Google's dominance in schools.

Reich said the report's findings indicate these problems cannot realistically be solved by individual teachers or school policies alone. "These are things that technology companies have to fix," he said. Google has not indicated it plans to allow users to disable AI search features, instead pointing to a "web" filter option accessible through submenus after searches are performed.

What remains unclear is whether the report will prompt regulatory action before schools complete their fall technology refresh cycles, when many districts finalize purchases of devices and platforms for the coming academic year.