Recently I have started to see more and more AI generated content on my timeline. I am pretty sure all of you have already seen OpenAI’s text-to-video generator Sora and yeah at the first glance this looks too good to be true. And in reality it is too good to be true. Because it is not true. We can see that one of the most viral videos generated by AI is pretty much a copy of a stock video. All these videos surfacing the web has one purpose. They are here to create hype. They are pretty good at it too. Now everyone thinks that we are all doomed but in my opinion all this hype is completely baseless.
AI has so many flaws and weaknesses that nobody talks about. I will list 4 reasons on why I think generative AI as in it’s current state is not sustainable and will die.
All these reasons will require AI to be regulated and when that happens that will be the end of generative AI.
I will go over each one of these reasons and explain them in depth. Let’s start
AI is data hungry therefore prone to stagnate
Ai needs real data created by real humans to be trained on. And the real data is finite. This makes AI models dependent on the creativity of people. It took years and millions of data samples to get to get to this point. Ai is extremely data hungry. It needs a huge amount of human-curated clean data to be able to produce what can be considered as a good product. And AI can only work with the type of data it is trained on. If you train an AI model on paintings of Van Gogh. It will be able copy his style and produce similar art pieces. Nothing unique per se. But just a pathetic attempt at imitating real art. This simply means that AI can’t imagine. It can only copy common patterns.