With the boom that ChatGPT is currently experiencing, more and more voices are pointing out that this chatbot or those to come that are able to improve what this OpenAI development already offers, they have the potential to take down web search engines which are still the main source of information search for the vast majority. And among these opinions we find equally qualified voices like Paul Buccheit, the creator of Gmail and a great expert on Google and the Internet.

So it’s not surprising that Google is closely watching everything that develops in the pair formed by Microsoft and OpenAI, and that those from Redmond stepped on the accelerator with regard to the implementation of ChatGPT functions, more specifically, a version designed specifically for this purpose in Microsoft Bing, your web browser. It seems that we are in for a paradigm shift in how we search for information on the Internet, and that artificial intelligence has a lot to say on this point.

Now, and we’ve mentioned it on previous occasions: ChatGPT, when you don’t know the right answer to a question, he has a bad habit of inventing. We already told you about it in this test, and since then I have found more invented answers using the service. Not on important topics thankfully, but made up and of course wrong, like when I asked him the best strategy for an idle game I’ve been playing for years and he told me he made up the game mechanics.

I highlight the problem of errors because of the news that we can read on Blu Radio Colombia, which informs about it a judge used ChatGPT to consult the legal framework for a court case. Judge Juan Manuel Padilla in connection with a trial involving a minor diagnosed with Autism Spectrum Disorder (ASD). Justice had to decide whether the child is exempted from the obligation to cover medical expenses including transport, or whether their guardians should instead take care of them.

“Judges must be aware of developments in justice and technology. Since the pandemic, in Colombia we have started to implement technology in the courts and this is a huge window, today it could be ChatGPT, but later there may be other tools that judges will use. This artificial intelligence helps us to create very understandable sentences, with good writing and so on.“, confirms the judge according to what we can read in the Colombian media.

I don’t know if this will be the first time something like this has happened or just if it is the first time it has been published (at this point it should be noted that the legal framework of the country expressly allows the use of this type of tool for this purpose), but without a doubt represents a very interesting milestone, but at the same time raises some suspicionsjust because of what I said at the beginning, ChatGPT’s bad habit of making up answers.

Of course, I take it for granted that the judicial professionals involved in this process, starting with Judge Padilla, somehow verified that the answer provided by ChatGPT was correct. If this is the case (and I repeat, I take it for granted that it was), we may be facing a precedent that we will no doubt soon see reproduced here and there. Now I’m holding my thumbs tight enough to break my phalanges waiting for verification processes to be fine-tuned to the max Well, otherwise it’s very scary to think about what we could be exposing ourselves to.