AI in 2027: Is it the year I become an ant?

AI is overtaking human intelligence in mathematics. Could it happen in geology?

I’ve been using AI tools quite a lot recently.

They have improved vastly over the past two years and are now particularly useful for coding. I’ve been vibe coding for the research work I’m doing with Bruce Hobbs and the results have been nothing short of astounding.

Only two years ago I had to rely on Bruce to generate 2D simulations using expensive, clunky software for our 2025 paper. Now I can do this in full 3D with code that AI writes in seconds, ready for me to play with interactively. AI has effectively replaced my human coder, who charged US$100 an hour and now does far more for US$100 a month.

AI tools are also very useful for literature synthesis.

AI quickly summarises what has been published. PhD students certainly have it easy these days. But when it comes to generating original ideas I don’t find it useful as it does not seem to have the ability to think laterally yet.

When you prod it with, what about this? it responds with, well, you are right and I didn’t consider that.

It will always try to force a round peg into a square hole. It will always consider the most popular theory/model and not consider less popular ideas. The scientific method, at least according to AI in 2026, is most of the time agreeing with the consensus, which isn’t following the scientific method at all.

But people in the know have been saying that AI will overtake human intelligence for many years now. In the past I doubted that it would. But in 2026 I am pretty convinced that it will and I think it will learn to think laterally.

AI surpassing human intelligence was predicted in the 2013 movie, Her.

In the movie the main character has conversations with his artificial intelligence, something we are already doing with AI in 2026. Eventually the intelligence and needs of the AI grow beyond those of a mortal human.

When I saw that movie I thought we might reach that point in about 50 years, well after I’m dead, so it didn’t really worry me. I wrote in 2018 that AI was nowhere near intelligent. But eight years on, in 2026, I get the feeling that we are actually at the tipping point.

It reminds me of my son growing up.

For the longest time my son, Umi, was much smaller than me. But for one moment when he was about 13 my size 8.5 shoes were the perfect fit for him, so he used to borrow my shoes.

That only lasted a couple of months.

Now his feet are size 12.

I knew from his growth pattern before he turned 10 that he was going to be at least 180cm tall (he’s 186cm). I marked that height on my wall but could scarcely believe it at the time. I feel AI development is a bit like this.

I feel that with AI tools in 2026 we’re at that brief point in time when my son’s feet were the same size as mine.

Mathematicians have recently had a taste of AI overtaking human capabilities as it is particularly good at coding and maths. AI has reportedly solved a previously unsolved mathematical problem, leaving mathematicians baffled. Some of the proofs are so confusing that they are still working through them, trying to understand them. Whether they are legitimate or complete garbage we won’t know for a while. But could it be that we’ve reached the point where AI is overtaking human intelligence, at least in mathematics?

Could this happen in geology?

Just imagine AI being able to reason logically, as it can now, but well beyond the intellectual ability of humans. In such a scenario we would be mere ants to it and completely irrelevant to discussions about things that really matter. The age-old problem of the origin of Witwatersrand gold would be trivial to such an AI. It might have already solved it back in June 2027 but we humans, with the intelligence of ants by comparison, might never be able to understand its logical solution.

So, in the very near future, we may be typing the following prompt into our favourite AI: Explain the origin of the Witwatersrand gold system as if you’re explaining it to an ant.

Or maybe not. But one thing’s for sure, it’s going to be a jungle out there.

*Jun Cowan is a structural geological consultant, specialising in interpreting mineral deposits at the deposit-scale and he runs practical structural workshops via StructuralGeology.com. He is the conceptual founder of Leapfrog software, which is now used by many international mining and mineral exploration companies. Leapfrog software resulted from private R&D collaboration undertaken by a joint venture between SRK Consulting Australasia, where Jun worked, and New Zealand company, ARANZ). Out of his home in Fremantle, Western Australia, he consults to mineral industry clients around the world and enjoys sharing his crazy ideas with his clients and with online colleagues.

 

Leave a Reply

Not registered? Register Now