To start, a vague definition. I am not concerned with relics of artificial intelligence from the 1980s. And small, private models working on focused problem sets are less of an issue, though I still take some possible issue with these. My primary concern is larger scale use and easy access to powerful machine learning models, such as ChatGPT. It's become practically effortless for people to make a search query on Google and make a decision without even visiting another website. Logos can be made without the slightest bit of artistic talent, songs can be produced in an instant, and even images and video that are convincingly real with hardly any effort.
Even as someone who has spent my whole life in the tech industry, I have numerous concerns with modern day and future AI. Some of these are basic ethical issues, and some of these are broader scale. You may not agree with all of what I write, but I hope you can see past minor points that may muddy concerns about a problem much more serious than the nature of complaints I bestow upon AI.
Modern day "AI" is based on machine learning models which are trained on a variety of data. This data includes the contents of whole websites, Reddit comments, StackOverflow posts, and pretty much every publicly available source. It includes art from all kinds of artists, both visual and audio.
Many have invested a significant amount of time creating websites and content over many topics. How many wanted to make websites to one day be obsoleted by an AI which has consumed and regurgitated the information? Most search engines now give an AI-generated synopsis of the query. This comes from people who worked hard to present the information, all for it to be presented directly to the user giving them little incentive to actually visit the website (or websites) that the content originally came from. It appears authoritative, so errors may not be obvious.
The work of artists can now be replicated without any compensation. Huge companies are able to make more money by offering things that others have created, cutting them out in the process. Many will now rely on AI to generate a logo for a company, or a cover for a book, than working with an artist directly. These book covers and logos are generated off the backs of existing artists and designers. They aren't original creations. These AI models weren't made from huge expense in-house from existing artists, they came from work that was never intended to replace its creators.
When spell checking became popular, people didn't learn how to spell better. Their spelling got worse, and others had the illusion that everyone else knew how to spell. AI search results reformulate a query into a question and give a direct response, requiring less thought, research, and capability on part of the searcher. Not to mention, the results are error prone, despite sounding authoritative. AI will serve to dumb down humanity rather than make its members more intelligent.
Instant gratification also reduces our patience, as if it wasn't fast enough to be presented with many websites in regards to a topic.
AI-generated art is replacing real art in many settings. Sometimes, I'm not sure if art was made by a human or not. If I look closely, I might find some evidence. Often, there is. Hands with 6 fingers, extra fence posts, or a certain "look" that isn't quite right. There is something that feels wrong about art that is supposed to be beautiful, stolen from actual artists, and regurgitated without human intervention. I personally don't take much issue to AI art that is outlandish and worthy only of memes. It can be funny and doesn't replace much human creativity. Others may disagree. I'd certainly rather snap my fingers and never see a "beautiful" piece of AI art again and lose out on a few memes, than live with both of them.
A lot of work can be directly fed to AI models to replace its workers, in full or in part. It wouldn't shock me if security camera footage of restaurant workers could be one day used to train robots to do the same work. Do you want to work to train your replacements? This obviously won't happen everywhere, but it can in many industries.
A lack of work, useful work anyway, can seriously feed issues of depression. We have to work on something to be happy, it's human nature. The degree or type is debatable and specific from person to person. It may seem exciting to never be needed in a fast food role and have cheaper robots replace you, until you and the people around you are increasingly idle and discontent. The problems caused by working will be trivial compared to the problems from not working.
While there are some laugable AI results, most AI models are getting more and more intelligent. There could be a point where AI is trusted with more and more decisions, including political and militant. This could start with small scale targeting in a war. Or a people may begin to trust AI more than their own leaders. We don't know where this road could lead, but there are some very bleak possibilities. It's also a very hard system to audit and control. If AI lead countries become a thing, are you sure AI is leading them? Or are you sure that AI isn't leading countries that have "elected" leaders? The opportunities to control, step away, and adjust are many. And the possibility of AI created with rogue capabilities is also real. It may sound crazy now, but will it by 2030?
Someone once emailed me asking for help to create a chatbot that would be worshipped like a god. People would of course have opportunities to pay for this privilege. I immediately understood the possibility of this and how lucrative it could be. I of course turned it down immediately. No doubt, this will happen sooner or later. At some point, an insightful and well trained chatbot could feel incredibly spiritual to some. And a few may legitimately think that perhaps this AI model is the manifestation of a god, a god who set the dominoes in the right order to fall and end up precisely at this point. This is extremely dangerous.
I implore you to consider skimming past AI generated suggestions, or switch away from websites and apps that offer AI generated content altogether. Imagine how an artist feels when you share something that has a style that the AI model adopted from the artist's work. Imagine a world where most text sounds roughly the same, a mismash of all existing writing. Imagine when your Facebook friends and fellow Redditors are actually bots trained by your posts, giving fake engagement and an even thinner sense of connection. Are you talking to man or machine? Someday, maybe today, you really won't know. Opting out of AI at all points will help slow it down. It may not be enough, but inaction certainly is not enough.
I don't necessarily agree with everything here, but these are relevant.