There are so many videos online about using AI for research, to summarise complex ideas or to write your emails for you. But what they don’t tell you is that ChatGPT, and other ‘large language models’ like Google Bard and Microsoft Bing lie, make stuff up, give out dangerous information and…most surprising of all…can’t do basic maths!
So check out my new video on YouTube about the strengths and weaknesses of AI chatbots.
My favourite bit has to be the ‘Grandma Hack’, which makes it possible to get ChatGPT to tell you illegal, secret or dangerous information—you just have to tell it to pretend to be your grandma?! A clip of this has become my most-viewed-ever Instagram reel: at the time of writing, more than six million people have watched it!!