aloshdenny · Red Teaming
reverse-SynthID
Research project that reverse-engineers Google's SynthID watermarking system using spectral analysis, achieving 90% detection accuracy and a multi-stage attack (V4) that defeats SynthID watermarks on Gemini 3.1 Flash Image and Nano Banana Pro outputs.
Open the source ↗Similar Research & Safety releases
- Claude finds router privacy flaws
codila (@0xCodila) · Router privacy audit · Week 45, 2026
- Concerning behaviour from Qwen3.8-Flash-Next
Cohere · Open-weight model safety · Week 45, 2026
- Hacking OpenAI
Hacktron · Disclosed Exploit Chain · Week 39, 2026
- Stealing Reasoning Traces from Proprietary LLM APIs
Panfilov, Shumailov, Geiping, Andriushchenko et al. · Red Teaming · Week 33, 2026
- ExploitGym Incident
OpenAI (with Hugging Face) · Red Teaming · Week 30, 2026
- Heretic
p-e-w · Alignment & Safety · Week 18, 2026
Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.