The hidden side of AI
In everyday life, our actions and words have serious consequences, or completely benign ones, with a whole spectrum in between. At all times, we must exercise our judgment and expertise, and take responsibility for what we say. Technologies evolve and advance; it is the role of professionals, engineers, and scientists to define their limits.
We must be serious and responsible, and verify what AI says, from beginning to end. Is the process of deconstructing a missing or flawed argument more laborious than thinking and searching for references? Certainly, for many of us.
AI presents a major pitfall: responsibility. In the worlds of healthcare, engineering, and science, verification and understanding are essential because discourse and responses demand accountability. Insurance companies, for instance, will seek scapegoats to make pay. We sift through a mountain of data, coded by non-experts—the Common Crawl—and emerge without any proof. AI feeds our intellect without providing any references. I believe that copyright should be paid otherwise. While we may claim that AI is improving, responsibility can never be evaded; an algorithm doesn’t live, it simply executes its instructions.
Unfortunately, irresponsible proponents have led people to believe they can heal themselves with AI, lawyers to invent legal precedents, and AI itself to invent diseases. They invented “Bixonimania,” a fake city, and a fake university. Some doctors believed it. Thanks to chatbots, this invented disease can be diagnosed. Fortunately, scientific experts are achieving remarkable feats by generating, for example, billions of 3D protein models—which, of course, need to be validated—using an AI tool, ESMFold2.
The tools that help us write, including dictionaries, are all referenced except for AI. When we use AI to develop our ideas, we lose the most important part of writing: reflection. I prefer those who learn to think rather than mere copycats. I myself experimented with AI’s ability to characterize writing using Vladimir Nabokov. The result is worth checking, the AI tells us, but you have to read it to form your own opinion.
Our writings, our works, our creations reflect who we are. What image of ourselves do we want to present? Copycats are easy to spot; you just have to meet them and confront them. AI shouldn’t just be a new way to cheat; it can help those who already understand and deceive those who want to learn.
– Jacques Gagnon, Eng., President and CEO of Imagem