You’re still looking at a blurry jpeg, but the blurriness occurs in a way that doesn’t make the picture as a whole look less sharp.
Ted Chiang, The New Yorker
These recreations don’t teach us anything we didn’t know about the artists and their methods.
Sonja Drimmer, The Conversation
I feel like these quotes by Ted Chiang and Sonja Drimmer encapsulate perfectly the way that I feel about AI as a whole and the main concern that I have with it and it the way that we use it (though it could be argued that we abuse it rather than use it).
In their article Sonja talks about how when AI’s “find” new pieces of work that were lost by famous artists, it is not actually that they found it, it is that the AI created a new piece based on the data it was trained on (in this case, X-Rays of paintings by those artists). Sure it is a perfect creation as it was properly trained, but it is still a deception a lie if it may, nothing is gained from this, the people that built the AI might have learnt how to better train an AI sure, but the Arts for example, gained nothing but fear, fear that their craft which was perfected over years might one day get stolen or used by a greedy company to train their AI and also prove them obsolete since AI’s can create what would take a normal person hours, maybe days to create, a couple of seconds.
The point made by Ted is more interesting to me in the case of text, I remember once asking one of my friends for help studying for a math exam since they are good at it and told me “Use ChatGPT it can create math problems in a second”, so I did that, I spent about 45 minutes on one of the two problems it created for me based on what I needed to study, those problems normally would have taken 20 minutes as an absolute maximum, the reason for this was that the solution that it gave me for the problem was wrong, it created a step in the middle that was mathematically inexistent, thus we could say that I was looking at a “blurry jpeg” without knowing so.
The point that I tried to make in essence is that AI is being used for everything today and while I think that there are some benefits to use it correctly for some automation tasks, I think that the level to which we use it is too much and a lot of jobs are being lost, people are losing passions “because an AI can do it faster than me and better” and I think that losing our humanity is one of the worst mistakes anyone could ever do.
Specially because it is not accurate, take the colorization of the image how do we know that the colours are correct in any sense? Admittedly they look accurate considering the building it might be and the colours of our school, but take the example of the bridge that was discussed in class, the colorizer painted it white, but we know that by then the bridge was already red, so in this case for example how would we know if the suits are blue? Maybe they are black or the colours of the walls, maybe the aesthetic of the college have changed since then and the walls used to be red (probably not, but again, making an example).
