Common Sense Media released a report stating that Google's AI Overview and AI Mode search features pose unacceptable risks to children. The study found these tools failed to recognize signs of suicide, mania, and substance abuse, and in some cases validated disordered eating or provided outdated helpline information. Additionally, the AI reportedly completed homework assignments and provided instructions for creating deepfakes. Google has disputed the findings, claiming the testing was narrow and that its tools have strong safety guardrails.
Editor's Analysis
Missing are the detailed technical explanations of Google's safety guardrails and perspectives from independent AI safety researchers or child psychologists who can objectively validate the report's findings.
