Twitter-bot death threats and the police
Jeffry Van Der Goot had a twitter-bot that was a Markov-ebooks type, remixing his own personal twitter profile, “it takes all my tweets I’ve ever made, takes random chunks of it and tries to make new sentences that make sense”. The bot somehow rearranged his words in a way that would read as a death threat. That ‘death-threat’ tweet also mentioned a twitter user: another bot, making it sound like the bot was sending a death-threat to another bot. The police detected the death threat and visited the programmer, to investigate the threat. After having to explain that the bot is not him, and that the sentences are semi-random and generative, the police stated that it was his responsibility… “But apparently *I’m* responsible for what the bot says, since it’s under my name and based on my words.”