Robots that read the world through cameras now lean on large vision-language models to interpret what they see and decide what to do next. These models handle images and text together, so any words that fall inside the camera frame become part of the input. A stop sign, a street name, a sticker on a wall. The overthinking behavior at the center of the attack Researchers at Michigan Technological University have shown that this reading … More
The post An AI overthinking attack can tie a robot up for over a minute appeared first on Help Net Security.
source & further reading
helpnetsecurity.com — original article
AI used to help plan the break-in, now it’s doing the break-in
“Context bombs” can frustrate AI-driven attacks, researchers found
The best defense against AI attacks turns out to be a skeptical human