AI models may be accidentally (and secretly) learning each other’s bad behaviors
www.nbcnews.com
A recent study is the latest to highlight a core AI safety concern: that the pace of development is outpacing humans’ ability to understand their own AI systems.