RIP Stanislav Petrov, "The man who saved the world." I used him several times in AI lectures as an example of why humans must be part of critical machine decision paths. An amazing story. npr.org/sections/thetw…
Aside from an example of human intuition, I also liked to point out that when Petrov was freer to speak after the fall of the USSR, he said his civilian training was essential. That a professional soldier might "act like a machine".
When he died around this date in 2017, I saw his obit and thought it might make a good "Day in History" item in a speech that week, but it had such an impact on the audience that I expanded it and used it a few more times. 
Now, when automated weapons systems are talked about a lot, and warned about by @NoelSharkey and @RespRobotics , it's a good story to keep in mind. AI isn't the Terminator, but we're good at finding ways to kill ourselves!
As I said in my talk today in Berlin, people's conception of AI is mostly from the movies. This makes us worried about wrong, unrealistic things (AGI, conscious killer robots, etc) & miss real concerns about keeping human society & well-being centered.
