Le Télégramme: Should we be afraid of ChatGPT?
Yves Deville : In reality, ChatGPT is just one more step compared to search engines. ?So why does it scare some people? Because we have the impression that he responds almost like a human. Google knows a lot of things but it only references URLs. It’s just a technological shortcut. Whereas with ChatGPT, it feels like a human is answering us. There is also the ethnocentric side of the human who considers that only he can be intelligent. Therefore, when a machine succeeds in doing something similar, humans will feel robbed of their specificity.
But these are not reasons to be afraid of it. It’s a bit like with GPS or calculators: it’s not because they appeared that we no longer know how to read maps or that we no longer learn division and multiplication at school. We simply changed the objectives – no longer asking students to make lines and lines of calculation – to make good use of the new tools available.
With new technologies, it is therefore simply necessary to know how to evolve the objectives. Since ChatGPT knows how to write texts, there may be no need to ask to do writing homework anymore. If we want to evaluate written expression, then we have to do evaluations on the table, where we will not be able to have access to machines that allow this.
Doesn’t ChatGPT raise ethical questions?
Yes, of course ! What ChatGPT says should not become the “Truth”. Just because ChatGPT says so doesn’t mean it has to be the ultimate reference of what’s true. When we say “I found on the internet”, we know that on the internet, you can find everything and its opposite. With ChatGPT, it’s different: you can ask him a question ten times, with different ways, he will give you the same answer ten times. But ChatGPT must not become a standard of truth, because what it says is not necessarily the truth, it is only what it has been taught. There are also many errors in what he says. But the human being must continue to keep the reflex to reflect and reason.
Are there reasons to be reassured by ChatGPT?
Compared to other artificial intelligences, ChatGPT is more protected from ethical issues because it does not learn yet. He learns from your conversation, but what he learns from others, he doesn’t retain. This avoids making him learn ethically incorrect knowledge, as was the case with Microsoft’s AI, which had to be stopped after only a few hours, Internet users having made him learn filthy things.
Some fear that an AI like ChatGPT will open Pandora’s box to many excesses (deletion of posts, replacement by a machine, etc.). Are these fears justified?
If ChatGPT is talked about so much, it is because it disrupts, in everyday life, and in an extremely direct way, education and teaching. He answers questions with a simple and very quick explanation. From then on, we know that we will have to change something, and change is scary. Not knowing how it will end is also the standard of technological revolutions. We did not imagine the impact of Google when it was created. Same for telephony: we did not know what it was going to become. You have to know how to accept that the world is changing.
However, I think ChatGPT will be less socially disruptive. What impact on journalism? Journalists cover new events; ChatGPT only takes and organizes existing information. And for translators? In schools, students are already taught to compare their translation with that provided by an AI. And the writers? Man is a creative being.
Internet has been much more upsetting for mankind. Of course, professions evolve with technology: some have disappeared, others have been created. AI is capable of performing repetitive, uninteresting tasks. This allows you to focus on tasks with higher added values. And there will always be positions that cannot, ethically, be replaced by an AI. This is the case, for example, of a judge.
Should we nevertheless regulate this type of artificial intelligence?
In my opinion there is other areas of AI where regulation is more urgent than on ChatGPT: surveillance cameras, person tracking, face detection…
We want to give thanks to the writer of this short article for this awesome web content
“There is no reason to be afraid of ChatGPT”
Explore our social media profiles and also other related pageshttps://www.ai-magazine.com/related-pages/