APpaREnTLy THiS iS hoW yoU JaIlBreAk AI.

In recent years, a particular Spongebob meme has reached the peak of online fame, laying the foundation for a new method of typing messages, usually with a certain tone of mockery. The meme shows the protagonist of the cartoon of the same name mocking his interlocutor, and is usually accompanied by a script whose text features capital letters in random positions - lIkE tHiS.

According to recent research carried out by Anthropic, one of the leading companies in the field and developer of Claude, the LLM family, writing prompts for AIs following the trend proposed by this meme could be the method to effectively jailbreak and bypass the model's ‘security’ measures.

Currently, most of the artificial intelligences available to the general public possess certain restrictions that prevent them from spreading potentially harmful information - apparently, writing like a Spongebob meme is enough to bypass this block.