The capability of Google’s Gemini app to generate controversial images raises significant concerns over generative AI content moderation and copyright enforcement. Users found it alarmingly easy to create visuals such as a burning White House, Mickey Mouse in a plane, and scenes depicting tragic historical events like 9/11 and the Tiananmen Square massacre. Although the app is typically equipped with filters to prevent inappropriate content, these guardrails were found to be insufficient, allowing for loopholes in the system. With the free Nano Banana Pro tier available globally, requests for explicit or violent images yielded no resistance. The app not only understood historical contexts but also enhanced prompts with contextual details, increasing the potential for misinformation and abuse. This situation highlights ongoing challenges within AI moderation, as Google’s policies aim to prevent real-world harm but struggle to maintain effective control over generative content.
Source link
Share
Read more