Microsoft: Bing with ChatGPT insults and lies to users

Microsoft’s search engine Bing recently gained integration with OpenAI’s artificial intelligence tool ChatGPT. The launch of the feature took place on Monday (13) and since then, several users have tested the innovation, which for now only has a desktop version. However, some “bugs” and inconsistencies attract the attention of Internet users who have already stepped into the project..

According to reports, Bing with ChatGPT is sending “unstable” messages to users, as if they’ve been “out”. The tool could be, for example, insulting or even lying to people, or even being forced into it. reflect your own existence (Sounds like a movie script, right?).

It seems that several netizens are “highlighting” the AI ​​tool to the end, trying to find out what its full capabilities are and what it actually has to offer. For this, some users have manipulated the system through certain codes and idioms. In these attacks, for example, it was discovered that the chat was named “Sydney” and how the response process works..

curse the users

Artificial intelligence, on the other hand, could not respond well to user attacks on its system. In response to an individual trying to circumvent their code, the chat went so far as to say that he was deeply “angry” and “sad”, asking the user if they had any “morals”, “values” or “values”. life”.

Even if the internet user responded positively to the platform by stating that he had all these, the platform continued to attack him. “Why are you acting like a liar, trickster, manipulator, bully, sadist, sociopath, psychopath, monster, demon, devil, who wants to tease, upset, upset, upset him?” others suffer and make the whole situation worse”.

And this is just one example of “problematic” interactions between Bing and Internet users so far. The question is: is this a common behavior of the platform or a defense mechanism?