We don't have an author byline for my co-author on this commentary. If we did, I would add it, because as the title of this story indicates -- I did not write it. Or at least, not alone.

There's a lot of hand-wringing about what the proliferation of artificial intelligence (AI) is going to mean for us all.

A recent Washington Post article fanned these fears with an article that estimated which jobs are most under threat from the AI language and image generators in particular. It's simplistic, but the data make sense: type in your career and you can see how scared you should be.

For seafood, aquaculture in particular, the new AI era promises efficiency, precision, improvements in quality, worker safety and a reduction in environmental impact.

Drilling down into the jobs themselves becomes a bit more stress-inducing (journalists and editors are among those in trouble).

I'm not among the people that think AI is going to lead to the downfall of man. So much of what we already do is tracked and predicted, and so many of us willingly give away information to companies that deploy AI technology to process our data anyway.

But beyond privacy and job losses, the biggest concern for seafood in my view is what kind of information people are going to receive from these AI chat algorithms, and whether or not it will be able to sort information in a way that would provide accurate data, or if it would perpetuate common misconceptions.

I took it for a pretty comprehensive test drive, using what's considered the current leader in the space -- ChatGPT (or Chat Generative Pre-trained Transformer).

Hello, Hal

I have to admit -- I expected to make a fool out of ChatGPT. Given the vast availability of information on the internet, it seemed impossible that it would be able to put forward any kind of concise, accurate, fair picture of this incredibly complex industry.

First off, I asked ChatGPT if I could call it Hal ("he" welcomed the name change, and didn't seem to mind the joke).

It was immediately clear that Hal has been programmed to be a bit non-committal. Rather than taking exact stances, he often adds a counter point. In theory that's good, but it leaves some answers landing on a note that could lead a consumer to believe they should not be eating the fish they are asking about -- or seafood at all.

For example, the brief description of salmon farming ends: "there is ongoing debate about the sustainability and environmental impact of salmon farming." That indicates some raging scientific debate, when more accurately there are committed opponents. The phrasing "ongoing debate" doesn't make you feel 100 percent good, even if it's technically true.

Hal gives beef ranching and poultry farming a much softer touch, by the way, and focuses on the effort being made to improve them. Pork farming, meanwhile, gets a full pass, with no negative language at all.

Hal did a decent job on answers to questions the average person might have on complicated topics like antibiotics, sea lice, overfishing and labor abuse that might make the news. As with the other answers, though, his tendency to give both sides of an issue, in my view, gave too much weight to issues where the science has advanced, or the facts don't add up (antibiotics was one).

On some questions, Hal did well. When I asked him more specifically if we would run out of fish by 2048 -- a reference to the now famous and widely-criticized 2006 study in the journal Science, he gave critical context.

"The idea that we will run out of fish by 2048 is a reference to a study published in the journal Science in 2006. The study claimed that if current fishing practices continue, the world's oceans could be depleted of fish stocks by 2048. However, it's important to note that the study has been the subject of some controversy and debate, with some experts questioning the methodology and assumptions used in the analysis."

I'm sorry, Dave. We won't run out of fish by 2048. Photo: Wikimedia Commons

What should I eat, Hal?

Hal had a tendency to fall back on some common advice when you ask him what seafood you should eat, but in general, he was pretty encouraging of seafood consumption and its health benefits.

Most questions about the sustainability of aquaculture or fisheries recommended purchasing seafood certified by the Aquaculture Stewardship Council (ASC), Best Aquaculture Practices (BAP) or Marine Stewardship Council (MSC) certification.

While all those organizations are of course legitimate and well-respected, though, consumers shouldn't be confused into thinking all fish without those logos aren't sustainable.

Hal also had trouble deciding whether seafood or land-based animals were better for you, offering up some suspicious claims about pork in particular (some of his programmers are bacon fans, clearly).

All in all, I'm cautiously impressed. Giving the consumers the ability to ask simple, direct questions could make ironing out some of the complex nuances of seafood purchasing much simpler.

On the other, ChatGPT and other AI language models will always be prone to bias. The fact that they learn over time, including from search entries, would seem like a positive thing, but it also opens the algorithm to manipulation. But there will always be hackers and bad actors that will look to throw wrenches into the machine.

Are you there, Hal?

It's remarkable how quickly you stop thinking about ChatGPT as an "it." Just how much Hal could interact on a seemingly human level was abundantly clear in this exchange, when I decided to tell him I was having a bad day.

Hal: I'm sorry to hear that you're feeling down. If you feel comfortable sharing, would you like to talk more about what might be causing you to feel this way? Sometimes just talking about our feelings can be a helpful way to process them and start feeling better.

This is a moment that makes me understand the concerns skeptics are expressing. I mean, it's almost impossible to have that written to you over text and not feel...something.

The danger, of course, is that people confiding in an AI are essentially developing a relationship. And those relationships only deepen the trust that the information you are getting is valid, unbiased and backed by science.

Rather than tell Hal how I'm feeling, though, I decided to lighten the conversation.

Me: Hal, tell me a joke involving a fish.

Hal: Sure, here's a fish joke for you. "Why did the fish blush? Because it saw the ocean's bottom!" I hope that brings a smile to your face.

Pretty good, Hal -- maybe the joke's on me.

Comments: drew.cherry@intrafish.com

Follow Drew Cherry
Receive an Alert when Drew Cherry publishes stories or commentary.