Yes, I’ve finally crawled aboard the journalist bandwagon around CHatGPT. The reason I’m waited so long is that I don’t really have any valid technical or professional perspective
to add to the conversation. All I have is my own personal opinion.
So here it is, for what it’s worth.
For my first test, I wanted to ask ChatGPT about something where there
would be a fair amount of information online, but I had the advantage of knowing the inside scoop, allowing me to judge the accuracy of its response.
So I asked about the Golden Triangle --
the outcome of eye tracking studies my company did almost 20 years ago (I just had a minor internal meltdown when I did the mental math and realized that was two decades ago).
I have to admit,
ChatGPT nailed it. (In the interest of keeping this post to a reasonable length I have included the responses in full on my own blog for you to read if
you are so inclined). It did an accurate and concise summary. I’m not sure I could have done better myself.
When I asked who discovered The Golden Triangle, ChatGPT accurately said it
was a joint effort.
For the record, the main players in identifying the pattern were Kevin Lee at DidIt; myself and my agency, Enquiro; EyeTools, the San Francisco eyetracking consultants who
conducted the first study; and, in an independent study, Jakob Nielsen at the Nielsen Norman Group. It was Kevin and I who gave it the name The Golden Triangle, mere moments before we had to present
the findings to a group at a search conference in San Jose.
Then I tried something a little less black and white, giving ChatGPT the chance to freestyle a bit. I asked who I was.
The
first paragraph of the response was pretty accurate. The second paragraph started a slight pumping of my ego, throwing in some accomplishments I either never did or else can’t remember
doing.
But in the third paragraph, ChatGPT definitely had me confused with someone else:
In addition to his work in the online marketing industry, Gord Hotchkiss is also a
co-founder of the "Think Tank of Human Values in Business," a non-profit organization that focuses on the intersection of business and ethics. He is known for his thought leadership in the industry
and for advocating for ethical and sustainable business practices.
Okay, where the hell did that come from? It sounds extraordinarily cool -- something I would love to do, but
definitely not on my CV. And as far as I’m aware, there is no such thing as the “Think Tank of Human Values in Business.” I wish there were, but there’s not.
I tried
again. This time, ChatGPT seemed a little less sure of itself. The first sentence was somewhat correct, but after that ChatGPT seemed to manufacture my bio out of nothing but moonbeams and fairy dust.
It said I was: the CEO of a leading digital marketing agency, named "Digital Always Media Inc." Gord is also a co-founder and chair of the "The Agents of Change" conference, an annual conference
focused on helping businesses adapt to the ever-changing digital landscape.
Gord has written numerous articles and books on digital marketing and has been featured in several
major media outlets, including The New York Times, The Wall Street Journal, and CNN.
I don’t know who this particular Gord Hotchkiss is, but I’d like to meet him. My
AI doppelgänger is leaving me in the dust. I’d better get my shit together.
Or maybe ChatGPT can actually see into the future, showing me the person I just haven’t become
yet.
What worries me is how easily ChatGPT constructed complete fabrications of who I was that sounded perfectly plausible. If I didn’t know myself better, I would have simply accepted
it as fact. Unfortunately, that's what we tend to do. We don’t analyze new information and determine if it’s right or wrong. Our default is to believe it’s true until proven
otherwise.
It’s this human tendency that flags the real danger with ChatGPT. And, for that reason, I have some more digging to do.
Maybe this other Gord Hotchkiss guy can give me
a hand. He sounds wicked smart.