Large language models (LLMs) like GPT-4 can identify a person’s age, location, gender and income with up to 85 per cent accuracy simply by analysing their posts on social media.
But the AIs also picked up on subtler cues, like location-specific slang, and could estimate a salary range from a user’s profession and location.
Reference:
arXiv DOI: 10.48550/arXiv.2310.07298
I tried asking Bing to make an assumption about who I am based on my Reddit account and wrote a nonsense made up story. Maybe I could have phrased it better?
This is basically what an LLM does, making up stories that might seem correct.
It’s statistics, basically.
People have to remember that, when they think it’s an all in one solution. AI is very powerful, but comes with realistic limitations.
Confidently making up stories.