If you’re a daily LinkedIN user like me you probably see no fewer than 20 “AI Pro-Tip” posts a day. Here’s another AI post but no pro-tips here. Just some things I’ve been noodling over.
I continue to be blown away by the capabilities that I’ve been experimenting with. Last week I read that the GPT-3 engine passed the MBA exam given by a Wharton Business School professor. The bot scored a B, which while not perfect, shows it’s mind blowing capability to do the work of an MBA grad.
For those of you that aren’t familiar with what exactly the “GPT-3” engine is, here’s the description I pulled from Google: GPT-3 stands for Generative Pretrained Transformer 3 and it is a powerful language processing tool developed by OpenAI. It is a type of artificial intelligence (AI) that is designed to understand and generate natural human language.
So get this: The next iteration of this engine, GPT-4, is nearing release and has 170 TRILLION parameters compared to GPT-3’s measly 175 billion parameters.
All of this has me sitting here wondering how this AI will be used in the research industry. How will it help us? How will it hurt us? Here are a few things that are top-of-mind for me and while it’s still a bit early to tell, my gut is pretty good.
One of the positives is that AI could be used as an intelligent “synthetic sample” in the early stages of research to test hypotheses, or hypothetical personas, based on its ability to approximate data based on historical data.
One of the biggest negatives that come to mind here are that high stakes business decisions will be made based on the voice of AI and not the authentic voice of the customer and lacks the human emotion and nuance of the customer experience.
This could even happen unknowingly if these smarter AI bots are deployed to successfully and intelligently complete research surveys similar to the bot farms that have been a thorn in our collective sides for the past decade. If the GPT-3 bot passed an MBA exam, why should we believe that a red herring question or 5 red herring questions and 3 embedded captcha requirements could trip it up, exposing it as a fraudulent response??
Interesting times ahead and with data integrity HIGH on the priority list for our industry, we need to keep our fingers on the pulse of this rapidly emerging tech.
There’s a good saying that comes to mind;
“If you stay ready, you don’t have to get ready.” 🤖
Jay Tye | Chief Operations Officer
P.S. This musing was written by a human!