If one of my social media platforms was giving my number out to strangers, I would be a little bit more than puzzled.

But obviously this man is a bit more diplomatic than I am.

The man in question, journalist Rob Price, was added to a random WhatsApp group chat, and then fielded random questions thrown his way in Spanish.

When he finally asked why they were asking him questions and why they added him to the group chat, they showed him screenshots of his number being saved as “Meta AI.”

They were told that the number – his number – could be added to group chats as Meta AI.

“You can add me to a WhatsApp group as if I were just another contact. You only need to save my phone number.”

As a journalist, Price is somewhat used to random contacts, but this was different, since people believed they were contacting a bot.

He eventually decided the mistake must have occurred while the large language model that is MetaAI was training. His phone number is publicly available so he can receive tips.

All LLM’s are trained by essentially consuming large quantities of information, so be thinks it ate his phone number and then made a “misguided causal connection” that resulted in it thinking his number was its number.

A spokesman for Meta agrees with the theory.

The problem? Axel Springer, the parent company of the publication where Price works, doesn’t have an agreement with Meta that allows their AI to train on their content.

The issue seems to have been resolved.

Meta is probably hoping the whole thing will just blow over.

If you enjoyed that story, check out what happened when a guy gave ChatGPT $100 to make as money as possible, and it turned out exactly how you would expect.