The Rapid Rise of AI: How Artificial Intelligence “Learned” 40 IQ Points in Just One Year

May 5

I’ve never been more afraid of robots taking over the world as I am right in this moment. Just kidding, I’m not scared, just a little curious and honestly excited for it to happen.

Okay, let’s back up for a moment and take a second to think about what it really means for an AI to jump from an IQ of 96 to 136 in a single year. That’s not just an upgrade, it’s like watching a high school sophomore suddenly start solving complex astrophysics equations and casually writing Shakespearean sonnets on the side!

AI hasn’t just improved, it’s evolving at a pace that’s starting to blur the line between machine intelligence and human potential.

But before we panic or start prepping for the robot uprising (already done, don’t worry), let’s break down what’s actually happening, and what it means for all of us.

What exactly does an “AI IQ” mean?

Okay, so AI doesn’t technically have an IQ in the same way people do. When you see headlines about an AI’s “IQ,” it’s really a fancy shorthand for performance on benchmark tests that mimic human problem-solving. Not the same test they gave you in high school or anything.

A score of 96 is about the “average” range for a human. And let me just add a little note here about how I personally don’t believe an IQ test is a real test of intelligence, as it can’t measure a lot of factors that make people successful.
But a score of 136, that pushes into genius-level territory, ranking smarter than 98% of people.

It’s a flashy stat, but it all boils down to this, today’s AI can process, analyze, and synthesize information faster and better than ever before, and it’s learning how to do things we didn’t even think were possible five years ago!

What changed in the last year?

The leap from 96 to 136 wasn’t magic, it was driven by a few massive shifts in how AI is trained.

Bigger models have entered the picture. Researchers have thrown more data and more computational power at AI than ever before. Think of it like giving someone access to every library, textbook, and expert in the world, and then speeding up their brain by 10x. It’s really no wonder it made such a big leap in such a short amount of time.

Also, we started using some better data than what we were using prior. AI used to learn from pretty shallow data pools, but now it’s slurping up high-quality academic journals, scientific papers, and industry knowledge. It’s not just memorizing Wikipedia anymore (no hate on Wiki, I personally adore it), it’s absorbing peer-reviewed science.

There’s also the little bump of more sophisticated training techniques going on. Engineers are using reinforcement learning and human feedback loops to “coach” AI systems, nudging them toward better answers and penalizing mistakes.

All of this together is like the AI went from community college night classes to an intensive Ivy League bootcamp. No hate on the community college world, went to one myself and it was a great life experience, never been to an Ivy League though.

Check out my deep dive into how AI already outperforms doctors in analyzing certain medical images!!

Why does it matter that AI crossed this threshold?

So the meat and potatoes of all of this is really coming down to the fact that when AI crossed the equivalent of 130+ IQ, it entered a weird new space.

It can now reason and problem-solve in ways that were previously locked behind human intelligence, it can generate new knowledge, not just summarize existing facts, and it’s getting much better at teaching itself, lowering the need for human intervention.

That last one I just mentioned? That’s the real wild card here. Once AI can self-improve, its growth may stop being linear, and start looking exponential.

But don’t freak out (yet)

I know this sounds like we’re on the edge of Skynet. But the truth is a lot messier and slower.

Yes, AI is smarter, but it’s also still incredibly brittle. It can ace a math problem one second and totally misunderstand a simple question the next. It’s missing emotional intelligence, ethical reasoning, and contextual judgment, the things that make human intelligence so nuanced and difficult to recreate.

Think of it like a super-fast calculator that’s starting to figure out metaphors.

Cool, yes, absolutely. Did we make a human though? Not even a little bit.

What could AI with a 136 IQ actually do?

Here’s where it gets fun (and a little scary) for those of you who need more excitement in your lives.

In business this level of AI could outperform human analysts in detecting market patterns, optimizing logistics, or writing reports. Just be cautious of those hallucinations that keep cropping up in anything AI generated. This is why I, personally, stopped using it for articles and switched to editor and image-generator. If you work in data-heavy fields, AI isn’t coming for your job tomorrow, but it’s definitely becoming a co-worker!

In science though, we’re already seeing AI propose new protein structures for medicine, predict the next breakthrough materials, and generate hypotheses humans hadn’t even considered. In one creepy test AI can apparently tell the difference between men and women with almost 100% accuracy just by looking at their iris and we have no idea how.

Yes, AI can now draft stories, paint pictures, compose music, and mimic famous artistic styles. (Though whether it’s truly “creative” is still up for debate.) But, as I mentioned earlier, it’s fickle in how helpful it can really be. It’s vastly limited by a lot of hallucinations and other things that AI seems to make up data when it can’t find any.

And yes, there are risks as well to all of this. Deepfakes, misinformation, algorithmic bias, they’re all part of this package. Which is why researchers are trying their best to figure out how to align smarter AI with human values before it becomes a runaway train.

What’s next for AI?

Nobody really knows, which I know makes you feel really secure, but here are a few predictions making waves:

Instead of one giant AI doing everything, we’ll likely see more narrow AIs trained for specific industries (like legal research bots, medical diagnosis bots, and finance bots).

We’re entering an era where the best outcomes won’t be humans or AI, but humans with AI. The smartest people will be the ones who know how to use AI as a tool, not a replacement. Because AI isn’t replacing any of us anytime soon.

More regulations will most likely pop up as well. As AI gets smarter, expect governments to crack down harder on how it’s used, who controls it, and what data it’s trained on.

Global AI races will continue as countries are already pouring billions into AI development. This isn’t just a tech trend, it’s becoming a geopolitical arms race. The next “Arms Race” might be more of an “AI Race” (and not the kind that happened in Beijing).

Why I’m watching this closely

Watching AI gain 40 IQ points in a year feels like watching a toddler suddenly speak in full paragraphs and do calculus by kindergarten. It’s exciting, unnerving, and completely impossible to ignore for a lot of reasons.

Whether AI becomes our greatest collaborator or our biggest headache will depend on how we shape it, right now, while it’s still learning from us.