I can buy the idea that simple specific tasks like translation will be dramatically cut down by AI.

But even then – any serious legal situation (like a contract) is going to want a human in the loop to verify that the translation is actually correct. This will require actual translator skills.

AI art seems to basically only be viable when it can’t be identified as AI art. Which might not matter if the intention is to replace cheap graphic design work. But it’s certainly nowhere near developed enough to create anything more sophisticated, sophisticated enough to both read as human-made and have the imperfect artifacts of a human creator. A lot of the modern arts are also personality-driven, where the identity and publicity of the artist is a key part of their reception. There are relatively few totally anonymous artists.

Beyond these very specific examples, however, I don’t think it follows that all or most jobs are going to be replaced by an AI, for the reasons I already stated. You have to factor in the sociopolitical effects of technology on its adoption and spread, not merely the technical ones.

It's kinda hilarious to see "simple task ... like translation". If you are familiar with the history of the field, or if you remember what automated translation looked like even just 15 years ago, it should be obvious that it's not simple at all.

If it were simple, we wouldn't need neural nets for it - we'd just code the algorithm directly. Or, at least, we'd be able to explain exactly how they work by looking at the weights. But now that we have our Babelfish, we still don't know how it really works in details. This is ipso facto evidence that the task is very much not simple.

I use AI as a tool to make digital art but I don't make "AI Art".

Imperfection is not the problem with "AI Art". The problem is that it is really hard to not get the models to produce the same visual motifs and cliches. People can spot AI art so easy because of the motifs.

I think midjourney took this to another level with their human feedback. It became harder and harder to not produce the same visual motifs in the images to the point it is basically useless for me now.

> any serious legal situation (like a contract) is going to want a human in the loop to verify that the translation is actually correct

I hope you're right, but when I think about all those lawyers caught submitting unproofread LLM output to a judge... I'm not sure humankind is wise enough to avoid the slopification.

Isn't "But even then – any serious legal situation (like a contract) is going to want a human in the loop to verify that the translation is actually correct. This will require actual translator skills." only true if the false positive rate of the verifier is not much higher than the failure rate of the AI? At some point it's like asking a human to double check a calculator

> But even then – any serious legal situation (like a contract) is going to want a human in the loop to verify that the translation is actually correct.

The usual solution is to specify one language as binding, with that language taking priority if there turns out to be discrepancies between the multiple version.

You might still need humans in the loop for many things, but it can still have a profound effect if the work that used to be done by ten people can now be done by two or three. In the sectors that you mention, legal, graphic design, translation, that might be a conservative estimate.

There are bound to be all kinds of complicated sociopolitical effects, and as you say there is a backlash against obvious AI slop, but what about when teams of humans working with AI become more skillful at hiding that?