What it is
Robert Miles AI Safety is a YouTube channel run by Robert Miles, a science communicator who began by making AI videos for Computerphile before starting his own channel. It explains why advanced AI might be hard to control, and how researchers try to address that.
What it covers
Videos deal with controlling AI, deception, biosecurity, AI policy and protests, with titles such as "Keeping AI in a Box", "Why Smart AIs Cheat" and "Bio Risk is Not Like Cyber Risk". Several recent ones also push advocacy, pointing viewers to ActOnAI.org and AISafety.com. Uploads are irregular: a burst in July 2026, then single videos on 1 August, 7 August and 29 September 2026 (as of 30 September 2026).
In an interview dated 17 August 2022, The Inside View quotes the channel as having 100,000 subscribers, and says he aims for the best available introduction to each idea rather than frequent uploads. That figure is dated, not a current count.
Who it suits
People who want a plain-English route into AI alignment without reading papers, and who are happy to follow a single presenter's views. It is commentary, explanation and advocacy, not news reporting. The videos in the public feed are openly listed; the channel page itself was not read.
