Here I am, on my way home from Nuremberg after the Legal Revolution conference. It seems that trains and flights are my best moments for blogging.
I was mesmerized by one of the speeches (given by Christian Keim, Adobe Systems Europe VP and head of international legal affairs). Everybody knows Photoshop (or at least they should) and maybe has grasped the essentials of its use, but what I saw during the speech was on another level. Imagine a picture with one or two characters, and in real-time, you can change the scenario, the clothes, the hair, the dresses. And being incapable of noticing the difference. All of this because of generative AI.
I was excited and scared about its uses. Think about political campaigns, Hollywood movies, or just the crazy, hectic social media world most of us are living in. Then think about the fact that we’re not speaking about the future, but of today.
Nothing that I didn’t know, of course (it’s been a bunch of years since I’ve start dealing with AI-related stuff), but it seems to me that speaking about AI, practicing AI, and watching its multiple uses provoke different feelings. Which sometimes are not in alignment.
However, my main reflection this afternoon (and the reason behind the post) was not about responsible AI or possible threats, but about the concept of true and false.
Those Photoshop examples made me ask myself: What is a true picture? The one where teeth are artificially whitened, belly fat is cut, and physical imperfections are removed?
And even cutting out Photoshop and similar programs, are we sure what we look at is real? How about a decent amount of lipstick, mascara, and kajal that your girlfriend puts on herself before a romantic dinner? Does this make her more or less real?
Change the scenario from visual to audio, and it’s pretty much the same. You could use my voice for something I never said just out of a 15-second audio clip and a written script, or you could hear me singing on autotune and change my pitch with Melodyne.
My feeling is that our entire perception of true and false will radically change in the next few years. Someone says that newer generations will grow up without this concept, without critical sense, without searching for the truth. They’ll take everything as it is, without a necessity for labeling. Maybe it’s just a matter of adaptation. I don’t know.
What I know is that – for now – at least in Europe, the AI Act (still not active) requires labeling AI-generated content, and a bunch of companies are pushing for an anticipated implementation due to potentially dangerous uses like these ones. Moreover, terms and conditions of various AI platforms require a declaration regarding the use, even though most users don’t.
As of today, I still value imperfection.
Joe Walsh (one of the Eagles) once said that AI “can’t destroy a hotel room. It can’t throw a TV off the fifth floor into the pool and get right in the middle”. And you know, that’s the essence of rock n’ roll…
