AI knows the probability of words in a sentence or paragraph, and has a framework for information based on its data set, but cannot be relied upon for factual information. Numerous lawyers have lost cases because they use ChatGPT to create legal briefs that cite nonexistent decisions, for example.
You are viewing a single comment's thread from:
OK I've got a question for you, just because I thought it might be fun to ask!
What's the difference between generative AI and the predictive text I just used to create this response to your comment?
Thank you for contacting me about the position on your family members of your family members and the trains are you coming back to the car with the kids and you will be back in touch to see what you think that's OK and I had a lodger of a few hours ago I have to tolerate the whole house and the trains are a couple of days before and I had to drive to the car park and I kept the keys with my head emerged.Generative AI uses much more complex algorithm and is trained on a far wider data set. Predictive text is based on a relatively simple library, and often based on your own text input history. The former has a better basis for sentence structure and syntax, and tends to be readable and coherent, if not necessarily factual and accurate. The latter just guesses what word might come next, and the more you tap/iterate on it, the less coherent it gets.