AI-Edited Photos Are Polluting Citizen Science: Gemini Gave a Heron a Third Leg
kscottz · x · 2026-08-18
The eBird and Macaulay Library teams warn that citizen-science projects like eBird, iNaturalist and the Macaulay Library are seeing a rapid uptick in AI-manipulated and AI-generated photos, a threat recently highlighted in a Nature paper.
The problem is worse than it looks: a user's photo edited by Google Gemini to "improve" quality gave a Green Heron a third leg; more subtly, AI fabricates false feather detail to make images appear sharper. Many phones and editing apps now apply AI processing by default with no way to disable it, so people may not even realize their photos are altered. This erodes the scientific value of the archives, and the teams ask participants not to submit AI-edited or generated images.
More from Multimodal
- Testing Song-dynasty aesthetics in image models: local hits, unstable holistic judgment — sujingshen · 2026-08-18
- ComfyUI tool uses model hashes to auto-fix red missing-model nodes — demongatanjieu · 2026-08-18
- Seeking Workflow to Render Fortnite Screenshots as Photorealistic Images — drmarkway · 2026-08-18
- 4DGS enables hyper-realistic 3D videos in web browsers with WebXR support — Scobleizer · 2026-08-18
- Fixing Minimax H3's blurry outputs: 4MP generations and Flux.2 refiner workflows — Patient_Ratio4177 · 2026-08-18
- Training Minimax H3 character LoRAs on 16GB VRAM: settings and pitfalls — No_Date4828 · 2026-08-18