Unless I'm talking about something serious, like fantasy football, conversations with me tend to veer into the absurd. And so, as I was having breakfast with colleagues in Denver at CEDIA Expo, a ridiculous name for an imaginary business venture was born. I went back to the Show Daily room and told ChatGPT to create a logo for that business.
The prompt was 18 words long. AI created the logo in less than a minute. It wasn't perfect, of course, but it was good enough to email and presumably get chuckles from my fellow diners.
I'm getting more comfortable using AI, even if it's for not-so-serious projects. For example, I wanted to get into the fall spirit for some videoconferences, so I directed ChatGPT to create an animated background of a coffee shop on a rainy day that incorporated our SCN logo. It took two prompts to get it close, a third prompt to add "DIET COKE" to the menu for me, and a fourth prompt to add "TEA" to the menu to accommodate my co-workers across the pond. It's not animated, but nobody's perfect.
From an imaging standpoint, AI is getting better all the time. The headshot at the top of this column? Created for free with AI using two current photos. That's head and shoulders (pun intended) above the results we all laughed at in 2024. It was also able to make me look like I belong in a Ranklin/Bass special without me having to explain the visual style of Santa Claus Is Comin' to Town.
Yes, the imagery continues to impress, even if I still have copyright and intellectual property concerns. (Seriously, where did those components for the café background come from?) On the research side, I still have my doubts. I'm going to get some flack for writing this, but I view AI as my hungover research assistant, and I treat the results from it accordingly.
Why so harsh? Several months ago, I asked ChatGPT to tell me about Mark J. Pescatore. It told me I was an industry journalist who writes for a different Future publication (not SCN) and one of my competitors (for which I have never written). ChatGPT also made the cardinal sin of declaring I was not the funniest writer in the Pro AV B2B space.
For this editorial, I tried again. While signed in as myself, I asked Claude, "Who is Mark Pescatore?" This time, I was included in a list of three people. I was the only one listed with my middle initial and was correctly identified as a "communications professional in Florida." It even listed my former PR gig—you know, the one I left back in 2021. When I asked again and included my middle initial, Claude reiterated my former position and mentioned my headline from LinkedIn. Not impressed.
When I asked ChatGPT this time around (also signed in as myself), it spat out a reasonably good bio, much of which is accessible right here on AVNetwork.com. And when I followed up immediately after that inquiry, ChatGPT quickly declared that I am now the funniest and best-looking writer in the Pro AV B2B space. (John Sciacca from Residential Systems was runner-up in both ludicrous queries, so congratulations to my colleague.)
In other words, much like a hungover research assistant, ChatGPT was trying to provide me with the answers I wanted to hear, accurate or not. In the past, when I've challenged AI on certain subjects or offered alternative choices, it has been quick to change its predictive and patterned tune, like some sort of digital toady. Swell.
AI is punching above its weight with meeting summaries, brainstorming, and turning headshots into puppets. Some people even see AI as a teammate, not a tool. I'm not there yet—I don't feel the need to thank my hammer or my AI—and I'm not sure I'll ever get there, particularly in this profession. Still, it's progress. Now, what would I look like as a bounty hunter in the Mos Eisley Cantina … ?
