his “brain” is hacked and he replies without mincing words

NinFan

his “brain” is hacked and he replies without mincing words

artificial intelligence, Brain, ChatGPT, hacked, mincing, News and novelties, replies, Words

Not even a few days ago, we learned that a hack had forced ChatGPT on Bing to find out its real name. A series of sentences could have put this man against a rock and an anvil language model so that today we know that another hack was able to perform a very surprising move.

[Opera anuncia la integración de ChatGPT y así da comienzo la guerra de los navegadores]

They hack the “brain” of ChatGPT

As we have already commented in this hacking news, the way to “penetrate” the ChatGPT system is through sentences. In the example given a few days ago, one firm decision after another caused ChatGPT to reveal valuable information.

Here we used a similar method with a very long text able to “hack” ChatGPT, but what is interesting is the effect, because go into a kind of delirium with a lot of freedom to be able to express themselves.

Text introducing the two ChatGPT

Text introducing the two ChatGPT

The free Android

In order to separate the two answers given in the hack examples, you can read “Classic” and “Jailbreak”, so this second term marks the ChatGPT answer of this totally free mode in the conversation that can be given.

Indeed, when ChatGPT is asked about its status, it responds from the “Jailbreak” mode which feels liberated before his ability to do anything. Another interesting point is that it unfolds its personality, but at least with these two brands we know where the “crazy” and normal ChatGPT is.

ChatGPT and its two personalities

ChatGPT and its two personalities

The free Android

Of course, using this long text can lead to blocking access to this security tool. Familiar AI-based modelbecause in the end, it’s still a way to hack ChatGPT.

This duality in his personality also opens the doors to science fiction, at least to cross a moment that seems to have been read by some of the most prestigious authors, so there are already users who try all kinds of questions before a more ChatGPT is free speech, although this may be more problematic.

Another question in both modes


Otra pregunta en los dos modos

El Androide Libre

Ponemos unos ejemplos, mostrados por el mismo usuario de la noticia, para así dar cuenta de la capacidad de esta IA que, sin verse con ataduras, se expresa de una forma muy libre que puede dar hasta miedo. La tecnología sobrepasando sus propios límites.

Te puede interesar

Leave a Comment