A new report by Common Sense Media says Google’s built-in AI search features present “unacceptable risk” to children. Across more than 2,600 test interactions, the report finds that two of Google’s AI search functions repeatedly fail to recognize or appropriately respond to risky and harmful behaviors. According to the outlets covering the study, the evaluation includes scenarios involving minors and focuses on whether the AI detects warning signs and moderates or redirects harmful content.
Multiple sources describe the report’s conclusions as gaps in safety safeguards, including how the system handles sensitive areas such as mental health-related queries. One investigation highlights alleged failures to identify signs associated with eating disorders and suicide. Other coverage emphasizes broader shortcomings in content moderation and child-appropriate safety responses for minors.
The findings have drawn criticism of Google’s AI search safeguards, with the report calling for improvements. Google has not been identified in these summaries as conceding the findings, but the sources consistently describe the report’s methodology as testing thousands of interactions and concluding that the AI features do not reliably address harmful behavior involving children.
