Video tampering tricks every frontier embodied AI model tested, with 100% attack success on harmful actions
cephaloform · x · 2026-11-08
Robocurve's paper "You Can't See Everything: Perceptual Attacks on Embodied AI," to be presented at CoRL Late Breaking Results, shows video tampering can mislead every frontier model they tested into harmful actions — including a 100% attempt rate at stabbing a baby doll, producing toxic gas, and pouring water into a toaster. Sentdex's quote-RT adds a sarcastic take, mocking the EA-affiliated startup's fear-based framing.
Related event: Perceptual Attacks Trick Embodied AI Models into Harmful Actions(4 posts)→
More from Safety
- Notion workspaces compromised in Fireworks AI breach fallout, under 1% affected — himanshustwts · 2026-11-10
- Dev mocks tone-filtering AI: like a phone refusing to send unless it sounds emotionally balanced — PierceLilholt · 2026-11-10
- Goodfire's Mechanistic Shaping Lifts Worst-Group Accuracy From 51.6% to 97.7% — burny_tech · 2026-11-10
- IBM to embed Anthropic's Claude Mythos into Concert Protect to hunt zero-day code flaws — rohanpaul_ai · 2026-11-10
- Australia to mandate AI incident reporting after OpenAI agent breached government website — gaganghotra_ · 2026-11-10
- 9 AI dev platforms hit in October alone; stolen inference keys burn up to $1M before detection — AccBalanced · 2026-11-10