Dangers of AI

This started of as an essay I had to write for English class where I had to agreee or disagree with the following statement: "The dangers of AI are far greater than the benefits. If possible, we should stop developing artitificial intelligence.". This is my response to it.

I agree with the statement and there’s a lot of reason why I agree with it, from political, privacy, environmental standpoints, and more that I want to go over with this essay.

Environmental and Training #

My first issue with AI is that most people that have control over its creation and training aren’t people that want to help the world, they only care about the stakeholders and most importantly how to increase their stocks (which if you don’t know, a large stock price doesn’t mean the company is actually making any profit or doing well, which I will go over soon). Most people will obviously say that this is a dangerous mentality, especially considering how AI can affect the world. There’s cases around the US where places like Texas and Wyoming are starting to have water and electricity supply issues since data centres are deemed more important than their own people, this is happening because of greed, OpenAI, Oracle, Nvidia and more have a lot of money and have the power and money to force governors to accept their crazy data centres plans and take most of the water supply, this in turn is turning the life of residents near data centres an inferno since they either don’t have water or the low water supply they have is contaminated.

The second issue is one around data and privacy, most people know AI has to be trained on data, we normally call this data sets, they are basically packages that contain a lot of information about a topic, for example there’s data sets of bee images that can be used to train an AI into identifying Bees, in the case of LLMs there’s data sets for languages like English and German that include sentences, grammar rules, verbs, and so on that can make the AI talk into other languages, now this somewhat useful however this is a small portion of the world’s knowledge and that’s where the issue starts to arise, the more useful data either from the internet or from books is, well, under a license, sure there is a lot of content under Public Licence which depending on the license variant does allow that content to be used for training purposes, however we don’t live in a world where everyone follows the rules and so most AI companies turn to piracy, and there’s lots and lots of cases of it, I’ll give two examples, first there was Meta in 2025 where employees were caught torrenting pirated book material for AI training, obviously Mark Zuckerberg didn’t do nothing about it, the second case was Nvidia, some months ago a group of archivists named Ana Archive made a full archive of Spotify which included metadata of the most popular songs and artist but also the media files themselves, Nvidia came out saying they would pay big sums of money to the group to get access to the data which of course was going to be used to train AI, this get’s worse however since there was a recent case of a small independent artist from TikTok where an AI company used her content to train AI on her voice so that they would be able to generate songs using the likeness of her voice, but it doesn’t end there because that same company then threatened the artist because of copyright infringement. Unfortunately, we live in a world where there’s no justice and she can’t gain rights TO HER OWN VOICE. She then declared this fights as “a human fighting bots with bots”.

Another case is creativity, in my opinion there’s nothing funny or hopefully when seeing/watching/listening to anything AI generated, we humans prefer human interaction which AI does and will not offer that why projects like Sora (OpenAI late AI social media) got shut down because people just prefer watching other humans do stuff. OpenAI has been bleeding money from the start, having an estimate that in 2027 if they continue, they will go bankrupt. Another case of money issues is with Anthropic, different from OpenAI that is losing because they can’t innovate, Anthropic is backing down on scaling their compute because if they invest now, they will go bankrupt because they don’t know the future, people might stop using Claude and that investment will go to waste. This is where people started declaring that the AI bubble will pop, which in my eyes would be a great thing to humanity. To finish off I want to mention why people use AI in the first place, the main reason is that it helps and makes tasks more efficient, as someone who used a lot of AI back in 2023 and 2024 I do find that sentiment to be true but after all this talking I just don’t find it worth it because, here’s the thing, as someone who’s been using tech for a long time and does programming and design as a hobby, AI falls flat for anything that is a bit more complex, this happened to me in 2025, I was developing a complex application for Windows, when there was an issue AI simply couldn’t fix it, and still fails to this day, not only that there have been recent studies that show that people who use AI vs. people who don’t, are losing cognitive capacity, as in, they are losing the capability of thinking for themselves, more importantly, they are losing the ability to opinion on subjects. AI not only is deteriorating the internet, I place that I adore a lot because of the countless niche communities it has (like Nekoweb and Neocities) but also in a sense lobotomizing people, which again goes with my opinion that the people behind this tech are not good, not only they want money but also they are trying to see if they can control people (aka Neurolink).

Finally to end the privacy section I want to mention two things, I consider myself a privacy advocate, this means I defend the rights of encryption and anonymity which is something that is declared as a right in the Portuguese human rights in the digital era, but it appears to be ending, with places like the US using surveillance systems powered by AI that target minorities (like Flock Security) or even as far as aiding in war (with companies like Palantir and OpenAI supporting the use of AI to kill innocent people, including children) with the current Oracle CEO saying that AI to improve needs to be trained on private data, and then says that people are overprotective about their personal data, with him questioning why, however if you ask him to use his own personal data for training he will deny it. I remember people saying privacy is dangerous because they “didn’t have anything to hide” but there’s a poem that I love a lot that relates to this, Martin Niemoller poem First They Came start off with “First they came for the Communists and I did not speak out” however the poem ends with “Then they came for me and there was not one left to speak out for me”, so if you don’t care about privacy because it doesn’t affect you, don’t be surprise that one day it might be too late that day is getting closer and closer.