Why is Google still serving dodgy ads? (www.atomic14.com)

Google continues to serve misleading advertisements despite having advanced AI tools capable of detecting such content. A recent case highlighted an ad in the YouTube app that mimics a system alert on iPhones, designed to trick users into believing their device is malfunctioning. Despite multiple reports, Google’s review process initially approved the ad, claiming it did not violate their policies, which raises questions about the efficacy and thoroughness of human reviewers in catching deceptive content. The significance of this issue lies in the potential for user exploitation and the impact on Google's reputation as a responsible platform. Google's own AI model, Gemini, quickly classified the ad as disapproved, highlighting issues such as misleading design, deceptive claims, and fear-based tactics. This contradiction between automated detection and human oversight suggests that leveraging AI in the ad review process could greatly enhance accuracy and user safety, pointing to a need for Google to integrate its AI capabilities more effectively to combat deceptive advertising practices.