On my new website "AI Risk, Plainly" I explain AI safety concepts concisely and accessibly, including the evidence for and against: https://leonarddung.github.io/ai-risk-plainly/