To say that generative AI is controversial would be an understatement. What started as just some surreal wobbly pictures and incomprhensible audio has since blossomed into a whole industry of algorithmically-generated media, leaving programmers and artists behind in the wake. Many despise it's existence, others embrace it. But why?
Listed below are a variety of peoples arguments against this technology, sorted into both valid and invalid reasons. Though the anti-AI crowd might seem more noble than those capitalist executives who wish to use anything to push out another penny, the truth is significantly more nuanced, and is best understood if you wish to improve both society and yourself.
We shall start with the valid reasons:
The fact stands that you don't really need to study anything to produce AI content, for the sole reason that you didn't produce it, the AI did. As such, it lacks that personal touch that makes art so special. Similarly, using AI to write something means you didn't write it, and as such you didn't really express yourself properly. This is especially important in programming, the "vibe coding" trend has been shown to be unreliable and produce buggy software.
Similar to the above reason, media produced by AI is significantly less appealing for the sole reason that it just looks ugly. Confusing forms, mismatched aesthetics and bizarre anatomy are common throught AI generated images, and although the tech is getting better, it hasn't done much so far.
In order to create an AI model for generation, it first has to be trained, which involves using an algorithm to refine a linear equation such that it fits within a certain bounds. This algorithm often has to go through the equation thousands, if not millions of times, just to arrive at something half-decent. As such, it consumes significnat amounts of power, power that could be used in something else more worthwhile.
There have been reports from AI engineers that the training dataset contians private material, including thousands of peoples personal information. Even more concerning than that would be reports of illegal content in training datasets, such as child pornography. To say that's a red flag would be a massive understatement. Thankfully, not all AI models have this, but it is concerning nonetheless.
Artists, writers, actors, etc rely on their jobs in order to make money and survive. But corporations don't care about that, they only care about the bottom line. As such, they see no problem with firing these workers just so they can use a cheaper, lower quality option. There is, however, another side of this coin, which we will go into in the next section.
A perfectly valid argument at first glance, after all, nobody want's their rights violated. But here's the thing, intellectual property is not a good thing. Things like copyright, patents, trademarks are only neccesary under a capitalist society, the same society that sees the use of generative AI as having no problems whatsoever. And things like copyright are only used to inhibit peoples creativity, by treating an unreal concept as if it was physically scarce, it limits peoples freedoms, forcing them to pay for something that should logically be free for anyone. Even the creators of the copyrighted material aren't safe, there have been many cases of them sell their rights to somebody else and thus not legally being able to do anything with the very thing they created.
This "argument" is often used by artists, who themselves are capable of drawing good art. The thing is however, in order for them to draw well, they had to spend hours, days, even years of practice, and even then theres' no guarantee that they will truly get good at it. People have other more important things to do with their time, after all. Some people don't even have the ability to visualize things in their head, which is a major stunt in their ability to draw well. This argument similarly doesn't work with other hobbies, such as acting.
Similar to the above "argument", this one also falls apart when you consider how expensive commissions are. They can often go high up into the 50-100 dollar range for just a single picture, even more if you want multiple pictures or an animation. Many people are simply unable to afford that, myself included.
While it is true that training a model consumes significant amounts of power, actually generating using it is significantly less so. Since the model is simply a linear equation, solving it is light work for any modern computer, just go through the numbers and crunch them. GPUs in particular were designed for doing that very task, but you don't see anyone complain about them wasting power.
This is the one argument that I personally have the most problem with. While it is true that companies shouldn't replace workers who depend on money to survive, there is nonetheless a basic truth that people should internalize: automation is not a bad thing. It would be great if jobs didn't exist, and that everyone was free to pursue whatever they desire. The only reason people currently want jobs is because they need money. What this effectively means is that being anti-automation is always pro-capitalism, since it denies societal progress in favor of endless work. The opposite isn't true however, since it's great for companies to not have to pay for workers to do everything. The issue isn't automation, it's automation in the wrong hands. If we want society to progress, we need to have this automation be for the good of the people as the whole, not just a few in the upper class.
So, what can we take away from these points that have been raised? Me personally, I don't think generative AI should be thrown in the trash and never used again. There are still plenty of valid cases in spite of that. There are plenty of open-source models available for download, even ones with their database linked and listed. It could be used for inspiration, for example, to illustrate something that would otherwise be hard to find with a simple search. It could also be used for generating the background to a proper art piece, something that people are likely not to focus on very much.
It should, however, also require regulation in regards to corporations attempting to use it to make a profit. Things like using cheap AI clones of actors should be what need regulation, not just the technology itself. It's why online communities that simply ban all AI without exception feel futile and petty, they're simply inhibiting the people without getting at the actual problem.
Technology itself is value-neutral. No one piece of technology is purely good or evil, it's the way it's used that defines the morality imparted on it. Seeing the world in black-and-white is unhealthy, you require nuance and shades of gray if you wish to truly understand anything about yourself.