As I’m following the generative AI bandwagon, OpenAI has released GPT-4, which is said to be “more dependable, innovative, and equipped to manage significantly more sophisticated directives” in comparison to its previous GPT-3.5 and ChatGPT versions. Not only that, but, in a demonstration, the engineers at OpenAI noted that the enhanced version of ChatGPT had outperformed 90 per cent of humans in some of the toughest exams in the US, including the bar exam.
See the photo summarizing GPT-4’s performance on several difficult exams in various areas of study and industries:
https://twitter.com/emollick/status/1635700173946105856/photo/1
GPT-4 has also been enhanced to handle images as a basis for interaction. OpenAI has provided an example on their website where the chatbot is given an image of baking ingredients and asked what can be made with them.
It’s interesting (and a little terrifying) to think about the potential that generative AI has, given how quickly it is evolving. It’s only been around half a year since ChatGPT was released, and its subsequent versions have already learned to do and accomplish so many more feats that are getting closer to what human capability and intellect can do. I’d be interested in knowing just how far its abilities will grow in the legal industry, and to what extent, it can take over legal professionals’ jobs.