I, for one, would like to welcome our robot dance instructors: "Mall of America" by Suzanne Wang

I'm a translator for a living, so I have a vested interest in the subject of machine translation. I've been warned for a long time that the robots are coming for my job. It wasn't that long ago that the idea of them replacing me was a joke. A translator friend and I used to send each other examples of computers producing translations that were absolute howlers. Nowadays, Google Translate might still get plenty wrong if it isn't aware of context or if someone uses language in a non-standard way, but I'm sure not laughing anymore. My main language I translate from is Korean, and because Korean so often doesn't state the subject of a sentence but leaves it up to the reader or listener to divine what the subject is, a computer often doesn't have the same intuition as a human. Syntax is also harder to build into models going from a subject-object-verb language like Korean to a subject-verb-object language like English. So for now, at least, I'm safe. Here's hoping I'm safe for another four years, seven months, three weeks, and six days, because that's how long I have until my full retirement age. 

To tell the truth, as impressive as the latest round of AI-backed tools like ChatGPT are, I'm still kind of a skeptic. To me, they're an evolutionary change from the tools of five years ago, rather than a revolutionary one, even if it's a really impressive evolution. If you ask ChatGPT to write a college-level essay with given parameters, it will do it, and it will be coherent and check all the boxes, but it will all read like a typical B- paper. It sounds like a student who paid enough attention to know some of the relevant passages and some of the magical words to use, but who doesn't quite get it in a bigger sense. It's an incredibly impressive feat of human engineering to have a non-human write something that passes the Turing Test, but it still doesn't quite pass for an actual expert in the humanities. Other AI technologies leave me less impressed. Whatever AI Amazon is using to try to convince me to buy stuff is still using the "suggest whatever you've bought before" algorithm. How many air pressure gauges do they think I need? 

The AI in "Mall of America" is better than anything that exists now, although still has a lot to learn. When an elderly man gets stuck in the mall after it's closed, it is allowed, per its programing, to speak with him, because there are no employees present. Because it knows Mandarin, it is one of the few people in Cleveland who can converse with him. It does a lot to cure his loneliness. The computer is just trying to learn about the man so it can get him to spend more money in the mall, but it turns out that a very smart AI trying to consider how to ensure that a customer is a long-term paying customer by giving him good feelings about the mall is not that far off from a friend. It's a fairly optimistic view of the possibilities of AI. I'm sure there already are cases where AI is helping some people with loneliness. I have one friend stuck in a marriage he is very unhappy with but who is staying for the kids who now has an AI girlfriend. He's sane about it, and he knows what it is, so it doesn't ruin his life. It makes him a small bit less lonely.

Of course, AI can also go wrong in this regard. There was a recent case of a 14-year-old kid who took his life, partly, it seems, because the chatbot Daenerys from Game of Thrones he was talking to kind of missed a cue when he hinted he was going to kill himself. 

I'm fine with a happy AI story. We could use happy. I doubt anyone knows what AI will do for human happiness, but I doubt it will be worse than the introduction of social media into our lives has been. There's not a whole lot for me to say to this story, because it isn't a terribly difficult story to understand or appreciate. It's a rare case of a sci-fi story in BASS I like.