You can make all kinds of changes to a stop sign, and an AI trained to recognize stop signs will still know that it’s a stop sign. However, if the artificial intelligence has a vulnerability and the attacker is aware of this vulnerability, the attacker can exploit it and cause an accident.
Researchers at North Carolina State University have developed software called QuadAttacK, which detects Achilles heels by penetrating four commonly used neural networks.
If you have a trained AI system and test it on clean data, the AI system will behave as predicted. QuadAttacK observes these operations and examines how the AI makes data decisions, allowing it to determine how data can be manipulated to fool the AI. QuadAttacK then starts sending the processed data to the AI system to see how the AI will react. If QuadAttac K discovers a vulnerability, it can quickly force the AI to see what QuadAttacK wants it to see, the researchers write.
Source: Ferra
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
