My hot take on AI as someone who's more familiar with what the training side of things looks like than a lot of people here probably are is that most attempts to turn using it as a random person in your own life into a moral issue are like. Kind of stupid tbh. Because "talking to chatgpt is Morally Wrong and you should feel guilty for it" is an extremely radical position that very few people are going to accept, and it's largely based as far as I can tell on a combination of "treating AI data centers as uniquely* harmful and evil when they're not" and like, personal aggrievement at the idea of having your own creative works used as training data, which results in a position that comes across as wildly overblown, overly sensitive, borderline hysterical, and excessively focused on imagining harm being done to you personally.
(*this is the most important word in this sentence. Large data centers can cause serious problems for nearby communities. AI is not the only reason data centers get built. Treating AI data centers as inherently worse than other data centers because they're AI makes you seem like you don't really know what you're talking about. There ought to be strict regulations about any kind of large facility that uses water and outputs pollution, period, but especially when they're built near people.)
The reason you shouldn't use AI is because it sucks. There are a few things it's genuinely good for, but it's getting pushed as a tool for finding answers and solving problems, when that's just not what it was made to do or what it's good at. Generative AI models are not search engines and should not be treated or marketed as such. They don't know things. They're inexplicably not even very good at reliably finding information, despite "sort through thousands of web pages, rank them based on a set of programmed criteria, and return a list of genuinely helpful, high quality, and reliable links" being exactly the sort of thing a computer SHOULD be really good at.
It's just not good. Its responses don't sound good. Its writing style is clockable and intensely mediocre. Because it's not actually capable of any sort of thought, it doesn't draw conclusions so much as rephrase things. The art it creates is blandly uninteresting. The things it's being heavily pushed onto consumers for both casual and business for are the things it's arguably worst at. It genuinely can't even do math.
Like the issue with AI is not that it makes people stupid, or is MORE environmentally harmful than like, any other physical internet infrastructure, or materially harms artists in and of itself, or that idk you are surrendering your capacity for critical thought if you use it to draft an email template for work. The issue is:
1) it's being heavily marketed and pushed in a fundamentally dishonest way that encourages people to believe it has capabilities it does not have
2) business owners think they can use it to make more money by not having to properly compensate actual human people
Both of these issues stem from labor rights, consumer protection, and general business regulation issues.
