First, let me say I don’t know Keaton nor do I know if Keaton actually did this. Second, I love Olive Garden’s response to Keaton on Twitter.
It’s when Keaton responds that I started to grow a little tired of the joke. He says: “Send it my way, and I’ll pass it on to the bot. I must warn you, it doesn’t take notes well.”
I am making an assumption that Keaton is using something based on the OpenAI framework. Here’s the thing. The bot should take feedback well because ChatGPT takes feedback incredibly well.
When I engage with ChatGPT, I will ask it to write like it is a financial educator, a career services employee at a college, a CEO of a Fortune 500 company, and much more — and it generally gives me results that jive with that. But, when it doesn’t, I rephrase my response and ask it to try again.
Because, artificial intelligence is doing what I told it to do — as well as the human trainers that trained it have told it to do. The artificial intelligence we use today is predicting text and other output based on our inputs. If it returned garbage like what Keaton received above, while entertaining, it points to a problem with the input that the bot received or the way in which the bot was programmed. And, most importantly, with the end user’s interpretation of the data.
We finished watching Hulu’s original series Class of 09 which takes artificial intelligence deep into the world of crime fighting. When…