The DeepMind AGI Safety team clarifies AI X-risk.

https://www.lesswrong.com/posts/GctJD5oCDRxCspEaZ/clarifying-ai-x-risk