ChatGPT is also very good at creating misinformation and conspiracy theories


Vincent Mannessier

February 02, 2023 at 2:30 p.m.

6

chatGPT © Shutterstock

© Shutterstock

Among the multitude of possibilities offered by ChatGPT, there is one a little less brilliant than the others.

Conversational AI can indeed create conspiratorial content at will, on any subject, and almost instantly. Chat GPT, on many subjects, is one step ahead of its own creators who often have to react with a delay.

You too can ask Chat GPT to create your own personalized conspiracy theory!

Several American newspaper titles have conducted experiments on Chat GPT on one subject in particular: disinformation and conspiracy theories. And it turns out that, for this topic as for many others, conversational AI is very effective. Disinformation expert Jim Warren managed to get him to say that a school shooting was a set-up by the media and the government, and that the families of the victims were paid actors.

Chat GPT is already very often wrong when asked real questions, and it does so with a certainty that is sometimes difficult to question. But on top of that, the AI ​​was trained from existing content to be able to adapt to as many requests as possible. And when it comes to conspiracy theories on the internet, that’s really not what’s missing. No problem finding material then.

Open AI is aware of the problem

The creators of Chat GPT are obviously very enthusiastic about the performance, but also about the possibilities offered by their baby, but they are not completely naive either. Even before the tool reached the popularity it enjoys today, they had equipped their AI with a number of safeguards. Thus, the conversational AI will not give ideas to budding stalkers or aspiring terrorists. This is obviously a step in the right direction, but for the moment insufficient.

Open AI, the parent company of Chat GPT, has also made it clear that the next versions of AI will be better equipped to fight against this phenomenon. This is far from the first challenge posed by the latter and, if not dealt with effectively, professional misinformation and troll farms in general could soon experience, like the tech world in general, a wave of layoffs.

Source : Futuristic, Chicago Grandstand



Source link -99