I have to admit, I don’t like how AI is developing. I hate ChatGTP, and generative AI in general, I think it’s totally useless. Why do I need an AI that creates poems, pictures, drawings, stories, translations, when the human mind is perfectly good at that? Why is AI being used in war? Why isn’t it being used more in medical science? Why isn’t it being used more for long and boring tasks like bureaucracy? Or, more in general, for tasks that take away time we could spend doing something creative? Like cleaning the house, for example. I think the little vacuum cleaner robot is much more useful than chatGTP giving me inspiration for a poem (there’s Pinterest for that anyway).
I know that technically the human mind should be good at doing bureaucracy too. But do we really like the idea that it takes days for any bureaucratic thing and seconds for a painting? Are we saving time in the right things?
I remember some time ago I saw an article titled “where is AI that pays my taxes?”, or something on those lines. However, when I searched it on Google for this blog, I couldn’t find it. I only found some actual AI that helps with taxes and finance. Which means that now AI for taxes exists. And who knows, maybe it’s actually useful. Unfortunately, the most famous AI are also the most useless. And sometimes dangerous too.
The only AI that I am ok with is the one used in science: medical science, for instance. AI used in war is completely against my morals, cause anything that contributes to war is. But when AI is used to manipulate information, or to substitute (creative) human jobs, it’s probably even worse.
As a translator, I’m really scared of not being able to get a job because “AI can do it faster”. Faster, yes, better, definitely no. Have you ever encountered the AI dubbing voice when watching a video on YouTube that’s not in English? If not, you’re lucky. I usually watch videos in many languages other than English (Spanish, Dutch, Romanian, Italian, Korean…), and every time, the first thing that I hear is this AI voice trying to translate everything in English. Basically, automatic English dubbing.
This has so many problems. First of all, it ruins the audio quality. Secondly, it doesn’t even work properly. Technically, you’re supposed to be able to choose whether you want to listen to the dubbed audio or the original one. So if you click on a video that’s dubbed, you can click on the three dots to go to the settings of the video, go on “Audio track” and choose the origial audio. Unfortunately, especially in the YouTube shorts, it doesn’t work. You can click on the origial audio option five thousand times, and you will still hear the AI English voice. And don’t even get me started on the translations. The ones made from Korean don’t even sound like English. They’re just random words. This means that it’s so literal that it’s incomprehensible, despite the fact that the goal of this AI dubbing is to make every video accessible to everyone, regardless of the language(s) that they speak. Also, AI can’t distinguish between a song and people speaking normally. So it will dub songs as well, with the same robotic monotone voice.
The only good thing that this is doing is reminding me that sometimes I have to put my phone down.
The fact that now AI is being used for translation has made a lot of things worse. For example, book translators have much less time to translate books, and sometimes they’re forced to use AI and to review its work, rather than actually translating. As a result, sometimes the translations feel awkward to read, and there’s not much effort to make the book enjoyable for people from different cultures.
Same goes for movie translations. In Italy, movies usually get dubbed. Only certain cinemas sell movies in their original language. So translators translate the movie into Italian, and voice actors dub it. There is a lot of work to make the dubbing as accurate and less annoying as possible, and the fact that they use human voices, which are much more versatile than AI voices, helps in creating dubbing works that sometimes are even considered better than the original. Not only for the translations, but also for the work of the voice actors. They really pay attention to details, like the movements and the expressions of the original actors, their voice tones, and how the move their mouth, because the lip-syncing is the most crucial part of translating and dubbing a movie.
Maurizio Merluzzo is the most famous Italian voice actor, and probably the one who worked at the most movies. He said many times that, because of AI and other factors, things are much faster nowadays, and it always feels like there’s not enough time to make an actually good job in dubbing a movie. There’s not enough time to pay attention to the details like they used to. Nowadays, when does a movie come out counts more than how is it like. Part of this problem is also that Italy was always one of the last countries to see a famous movie because they had to dub it first, so when we got to see those movies sooner it felt like a dream. Unfortunately, what we see it’s definitely not the same quality.
Another thing that’s disturbing about generative AI is the way it’s being used to create drawings, pictures and videos. The Italian government has used AI to create drawings for propaganda (especially anti-LGBT propaganda) and the results were quite disturbing. AI is also used to create fake videos (deep fakes) and pictures that alter reality in a way we really don’t need right now. There were many of these videos during the war in Ukraine or the genocide in Palestine.
The worst thing about it is that the Meta apps use everyone’s pictures on social media to train the AI. Any photo posted, for example, on Instagram, can be used to feed the AI’s repertoire and create better fake photos. Having a private account won’t save you in this case. If you don’t want this to happen, you have to fill out an official form with a message saying that you do not give consent. Basically all the people I know have signed this form. And my question is: how many people in the world have signed this form? And a follow up: let’s say that they recieved these forms from more than 50% of the users, why did they still went on with the idea? Also, why do I have to sign a form to NOT give consent to this? Shouldn’t it be the opposite? Why is it the default to give consent to my photos being used to train AI to create deepfakes?
This is a reminder to sign that form so that your pictures on Instagram will not be used to feed AI.
Nowadays everyone has heard about the new proposal from the EU, the “chat control” law. In order to protect the children online, each single message that everybody sends on any app will be sent to another server and controlled before being approved and sent. Experts have already warned about the risk for the people’s privacy.
The idea of protecting children from any type of abuse online is great and I totally agree with it, but is the EU taking action in the right way? Is it really necessary to ruin the privacy of innocent people who just wanted to message their friends? Isn’t it better to invest in educating families about the usage of internet? Controlling every single message from every single person sent online everyday sounds like an unecessarily long and hard work. They’re just bypassing the problem.
So I hope that someday we will all collectively realize that we don’t need ChatGTP to search things, to study, or to create. We already have a lot for this and ChatGTP and AI in general doesn’t really give anything more. We need to understand that we are capable of doing those things, even if we’re slower than a robot. We don’t need AI doing art for us. We’re supposed to collaborate with it for the society’s good, which includes scientifical and other technical discoveries, but also saving time for creativity and scientifical/technical research. Letting AI replace or manipulate us will not bring us anything good.
https://edri.org/our-work/chat-control-what-is-actually-going-on
Recent Comments