Maybe Marty the Supermarket Robot could moonlight as a crimestopper.
If police were to arrest and jail you because some fortune-teller peered into a jar of Miracle Whip and announced that you were about to commit a crime, you would surely feel yourself ill-used.
But if a robot using “Artificial Intelligence” predicts it, well, that’s different (https://www.dailymail.co.uk/sciencetech/article-7287341/AI-experts-release-statement-slamming-predictive-policing-digitizing-stop-frisk.html). It’s scientific!
See, they want to use “pre-crime algorithms” to know who’s gonna be bad so that they can bust him before he does it. Never mind the critics out there, most of whom are described as scientists, who call the whole thing “useless” and warn that it might lead to “mass incarceration” of people who haven’t yet committed the crimes they were jailed for.
The critics point to statistics that show that really very few people go out and commit a new crime while awaiting trial for an earlier offense. Another arrest–well, it looks bad. It might make it really hard to get bail. Judges don’t like to grant bail to persons who have just been busted for yet another crime. You can see their point.
Meanwhile, listen carefully… There is no such thing as “Artificial Intelligence.” There is only whatever human intelligence, or lack thereof, that goes into programming the computer. Algorithms are human creations–and therefore eminently fallible.
Would it help if they programmed the robot to say “I’m sorry”?
(Ooh, ooh, I know! Why not just lock up everybody! Robots could guard them and keep them from escaping. Good idea?)