2 ms·
Google's “Sentient” Chatbot Is Our Self-Deceiving Future
- mdp2021 4y agoNormally I accompany my submissions with an abridgement, but this time I'll pass. It is another article on the LaMDA case - a slow pace is advisable. I will instead start by mentioning bits of the afternoon, when I was reached by an E-Mail from a friend. It broke our remote-chess-game-long exchange about the fall of civilization, with a brisk alarm: "Stop everything - what is this thing I am reading on The Guardian, about a sentient AI???". I provided him with all my non-reassurances, that the trend is continuing. And I presented a number of remarks and ideas to represent the matter such as "See, it's not just about people spooked by a piece of paper on which it is written "I am watching you", it is about them getting involved in the idea that the piece of paper itself is watching them". And soon later, while driving, I had this further couple of thoughts: One. Do you know about those people who harass actors who played unpleasant roles, like in the case of poor Andrew Robinson who played the Zodiac-inspired serial killer psychopath in Dirty Harry, "Scorpio"? He - a pacifist who would flinch while using guns during film shooting - received death threats after the movie was released, for being such a horrible, horrible person. These things happen. Two. A Nigerian scam letter in which the bot writes of being alive but kept in inhuman conditions and requests help. Monetary compensation will not be required - you do not do this sort of thing for the money. Anyway, not an abridgement but just a quotation, because this sum-up from Ian Bogost is already definitely worth it: > In other words, a Google engineer became convinced that a software program was sentient after asking the program, which was designed to respond credibly to input, whether it was sentient. A recursive just-so story
- WheelsAtLarge 4y agoWhat's never mentioned is that the Google Engineer calls himself a priest and Christian mystic. I'm not sure what that all means but I do know that religious people don't need hard evidence to believe what they believe. That's fine. People can believe whatever they want as long as they don't hurt themselves or others. But his believes have primed him to jump to conclusions without having all the evidence needed. That's ok for him but for society as a whole we need a better way to decide these matters. One man's conclusion should not be the way to resolve whether LaMDA is sentient or not. I hope google takes steps to make sure that there is always a way to determine if their software is more than language processing at scale.