Generative AI

(This is a re-write and extension off of an old BlueSky thread I wrote almost a year ago now. My thoughts haven't changed much in the overarching themes, but I want to ensure I'm not misunderstood, to reiterate my feelings for those in the back, and to back up some of my thoughts with sources where I can. This is by no means comprehensive - I'm just so tired of this shit, dude.)

Generative AI erodes Trust

"GenAI", in the way it has broadly been used, erodes trust. It should not be understated that in the time since I initially discussed this on BlueSky, "AI deepfakes" have become increasingly common in such a disturbing fashion - and that's just the most recent stuff. Before October, you had people who were faking to be artists just for a quick buck, or artists who seemed to have given in out of fear of being "left behind". It's always done behind the scenes, in a sneaky way because they gotta know they're doing something wrong, or at the very least that there's consequences for being caught.

(Quick addition here, as this was written before Sora 2 came out: Sora 2 makes this worse for the average person. That's all.)

Generative AI ruins Teamwork

Why work in a team when you can just have your buddies ChatGPT, Claude, etc. all do your writing, artwork, and programming all for you! Surely there's no problems that could come from that!

(There were, in fact, several problems that could come from that.)

Even if it all somehow came out perfectly (which is, at best, unlikely), you rob yourself of the chances to work with others on those projects. Yeah, maybe working with others doesn't turn out exactly perfectly to what you wanted, but that's the fun part: experimenting, learning, growing, and building out a project with a team. It's infuriating to hear from people who are so scared of the idea that they "might not be able to articulate" themselves correctly that they would rather hide away and either do nothing, or gamble with every single asset of their project in hopes it might be "good enough", rather than the best it could be.

You cannot "democratize" art by deliberately removing everyone else from the equation. That's a fucking oxymoron.

Generative AI is Not "Inevitable"

... In fact, survey says we're in a financial bubble because of it.

Not long ago, OpenAI (the one with the most market share) got a deal with SoftBank for it's last round of funding. They got $40B back in April, but a decent chunk of it is predicated on OpenAI going for-profit (which they can't really do), so $10B of that will become debt in the next couple months. Bear in mind, that OpenAI's been burning cash at an astounding rate, and it's only real competitor, Perplexity, isn't exactly doing much better from what I heard.

Companies are literally having to push GenAI on users, much to the chagrin of your local IT department. Anything you put into these tools is pretty much guaranteed to go right back into their datasets, so you bet that's a security risk, and anyone telling you it isn't either doesn't know what the fuck they're talking about, or are probably trying to sell you on it. If it was so good, why would you need to push it onto people like the fucking plague?

(Another addition after Sora 2 came out: Sora 2 makes the costs worse, if the briefest bits of posts by Sam Altman are anything to go by.)

"Ethical" Generative AI is a Myth

The tools have been proven, time and time again, to be built off of the stolen work of writers and artists.

(Even fucking GameFAQs wasn't spared, like- really? Don't you have anything better to do?)

There's not a debate about this fact. It's what encouraged tools like Glaze and Nightshade to be created, for protection and dataset poisoning, respectively. They're even working on a version of the tool that combines their effects together. People were forced to find solutions to fight back because the Tech Industry doesn't understand consent, but realistically if your artwork had already been uploaded for years it's likely already in a dataset somewhere, and there isn't much that can be done about it right now. Imagine how bad it is with the folks who're dead, and getting their work stolen for this junk (and this was a concern before the deepfakes started to become hard for the average person to catch). Nothing about this screams "ethical".

In Conclusion

... So no, I do not like Generative AI. I will not use it for my work. I do not want to see what you made with it, or how "funny" it is that it made a "mistake". I'm tired, and I just want this goddamn bubble to pop already so it can end already. The aftermath of that will be hard thanks to some bad decisions in our economy, but the longer this goes, the worse I fear it will be.

Thanks for reading.
- Chloe