Chatting AI & conservation: Trying to create God to solving real-world problems

I'm a Luddite at heart.

I've spent my life embedded in the outdoors, trying to be untethered to modern technology.

Reality is that even those who share my Luddite leanings are entangled in modern outdoors technology. Luddites fought, unsuccessfully, against modern machine in the 1800s. Frankly, modern technology enriches many outdoor experiences.

I can't imagine being without my outdoors apps: weather radar, Seek (for identifying the natural world), PictureThis (most useful plant identifier) and Merlin (bird ID). So I attended "Nature's New Ally: Can A.I. Fix What We Broke?" session at the biannual Wild Things Conference in Rosemont. The sold-out conference of some 2,500 was packed with more than 100 sessions.

Robb Telfer, Friends of the Illinois Nature Preserves program director and program coordinator for the Wild Things conference, emceed the AI session. Philip Willink, a researcher for the Illinois Natural History Survey, and ecologist/botanist Evan Barker, Pizzo Habitat Restoration and adjunct professor at DePaul on plant identification, did the pro-con presentations.

Telfer wondered about the definition of AI. I like NASA’s base definition: "Artificial intelligence refers to computer systems that can perform complex tasks normally done by human-reasoning, decision making, creating, etc." NASA then makes more precise definitions and distinctions from the related Machine Learning.

Willink went first from the pro side. Even those who oppose AI will admit there is potential for conservation. Though, Willink noted, "I'm getting a sense there's a lot of difference of opinion."

Then he pointed out, "Probably everyone in this room uses nature apps."

Bingo.

"I was interested in pushing the boundaries more," he said. "This is all good, but can it solve real-world problems? Can AI take the next step?"

He set a real world problem for ChatGPT. A culvert with a three-foot drop created a gap in fish movement. Illinois has 20-30,000 culverts. The fish above the culvert were isolated from the rest of the waterway. The question was how fish could swim up and down culvert.

He already knew the answer from encountering and solving the problem in real life. AI came up with a good answer.

"A streams biologist would take days to a couple weeks to solve it, AI was able to do it under five seconds," Willink said.

To improve the answer, he went back and asked the question in other ways to play around with the answer.

"AI asked about the target species," he said. "I mentioned sculpin and stickleback and answered it like a fish biologist would have answered."

Feedback was important for finding better answers in the future.

"While I was training AI, AI was training me to think like AI," Willink said. "The Pandora's Box of AI is opened and we are not putting it back in the box. We need to figure out where this is going and get there first.

"AI and Quantum computing is going to dramatically change things in 15-20 years."

He thinks there's much potential for good with prairies, but added, "You have to ask the right question within the right framework."

Barker made it plain he was on the con side.

"AI companies are trying to build God whether they realize it or not," he said. "Today is the worst AI will ever be. It will only get better. That's actually scary."

His concern is more with the people using AI.

"Agents are something you will hear a lot about in the next month," he said. "I worry more about what we do with AI than what it does to us."

He pointed out that Seek is already better than he is at answering questions on plants. Sites species lists are now available for anyone to see online.

He had a warning that had never crossed my mind, "Poaching will become extreme with enhanced AI."

He offered a sliver of hope, "The light at the end of the tunnel is that I am up here preaching to you about this."

But he also granted that naysayers like him have been wrong at other times with dire predictions, "I hope I am that naysayer guy who is wrong. Big thing for me is what we don't know."

Then he threw in a scenario more plausible than I care to admit, "Image scenes where AI becomes self aware. It is going to have a large bank of information. We've written all these sci-fi stories for it."

The Q&A was avid.

An audience member warned that the energy used to run AI "is literally going to boil lakes to cool the giant computers."

Another had a different take, "If we abandon AI, we will leave the battleground to those who do not share our values."