I heard someone say something the other day that stuck with me: “AI is smarter than everyone.”
At first, it sounds reasonable enough. Depending on what you ask it, artificial intelligence can explain quantum physics, write computer code, discuss Roman history, diagnose what is wrong with your Excel formula, help you understand a Supreme Court decision, and suggest what to make for dinner. Most of us are doing well if we can remember why we walked into the kitchen.
So, yes, compared with any one person, AI can certainly appear ridiculously intelligent.
But the more I thought about the statement, the more I wondered whether we are describing what AI actually is correctly.
Is artificial intelligence really smarter than everyone?
Or has AI simply been given access to what everyone already knows?
That distinction may sound like splitting hairs, but I think it changes the entire conversation.
What Does “Smarter” Actually Mean?
Before deciding whether AI is smarter than humanity, we probably need to decide what we mean by smart.
Knowledge and intelligence are often treated as interchangeable, but they are not. A person can know an impressive number of facts and still demonstrate absolutely terrible judgment. Anyone who has ever sat through a corporate meeting knows this phenomenon is alive and well.
Knowledge is information. Intelligence is what you are able to do with it.
A physician knows medicine. An engineer understands structures. A historian understands events and the forces surrounding them. A mechanic may hear an engine for five seconds and know exactly what is wrong with it. A successful business owner may look at three numbers on a financial statement and immediately recognize trouble that someone with an MBA and seventeen charts somehow missed.
Each possesses knowledge, but intelligence appears in how that knowledge is interpreted, connected, and applied.
Artificial intelligence complicates this because it has been trained on an extraordinary amount of human-created material. Books, research papers, websites, software, historical records, technical documentation, literature, scientific discoveries, arguments, explanations, examples, successes, failures and, presumably, enough terrible internet comments to qualify for combat pay.
No individual human could consume even a meaningful fraction of that information in one lifetime.
AI can work across enormous amounts of it in seconds.
That certainly looks like intelligence.
But it may also be something humanity has never had before: near-instant access to an enormous portion of our own accumulated knowledge.
Put Everybody Who Knows Something in One Room
This is where my brain went next.
Imagine that we could gather every person on Earth who knows something useful into one enormous room.
Not just Nobel Prize winners and university professors. Everyone.
Put the physicists next to the engineers. Bring in the physicians, mathematicians, historians, economists, psychologists, programmers, architects, writers and scientists.
But also bring in the farmer who has worked the same land for forty years. Bring the captain who has spent thirty years at sea. Bring the mechanic who can diagnose an engine by sound. Bring the nurse who has seen ten thousand patients. Bring the carpenter, electrician, machinist, chef, soldier, pilot, teacher and business owner.
Now you have billions of years of combined human experience standing in one place.
Would artificial intelligence still know more than that room?
Probably not.
In fact, humanity collectively almost certainly possesses vast amounts of knowledge that have never been written down, published, digitized or uploaded anywhere. There are things people simply know because they have done them thousands of times.
Researchers sometimes describe this as tacit knowledge, meaning knowledge that is difficult to completely transfer through words.
You can read every book ever written about sailing and still not possess the instincts of someone who has spent thirty years at sea.
You can read every leadership book on Amazon and still discover that leadership feels somewhat different when thirty employees are waiting for payroll and the company checking account is short.
Experience has a way of adding footnotes that textbooks forgot.
So our imaginary room probably contains more knowledge than the AI.
Unfortunately, our imaginary room also contains several billion people trying to talk at once.
And that is where AI wins.
AI May Not Know More. It May Connect Better.
Suppose we have a complicated problem involving logistics, psychology, engineering, economics and history.
Somewhere in our giant room are people who understand every piece of that problem.
The difficulty is finding them.
The logistics expert needs to locate the psychologist. The psychologist needs information from an economist. The economist realizes a historian studied something similar fifty years ago. The historian disagrees with another historian, so now we need a meeting.
Someone schedules the meeting for Tuesday.
Two people cannot attend Tuesday.
It gets moved to Thursday.
A consultant creates a 74-slide PowerPoint presentation.
Six weeks later, everyone agrees to form a committee to investigate the matter further.
Human civilization has entered the chat.
Artificial intelligence eliminates much of that friction.
Ask AI a question that crosses several disciplines and it can attempt to connect information from all of them immediately.
That is an enormously powerful ability, but I am not convinced it means AI is inherently more intelligent than humanity.
It may mean AI is dramatically better at connecting humanity’s intelligence.
And that is a different idea altogether.
We Have Been Building This for Thousands of Years
Once I started thinking about AI this way, it stopped looking quite as much like a sudden technological miracle.
Instead, it began looking like the latest development in something humans have been working on for thousands of years.
Before writing, human knowledge existed primarily inside human beings. When someone died, whatever they knew that had not been taught to another person could disappear with them.
Writing changed that.
Suddenly knowledge could survive the person who discovered it.
Libraries allowed that knowledge to accumulate.
The printing press allowed it to spread.
Modern telecommunications allowed information to move almost instantly.
The internet connected enormous amounts of information.
Search engines allowed us to find it.
And artificial intelligence may represent the next step: allowing all of that information to interact.
That progression matters.
For most of human history, the problem was preserving knowledge.
Later, the problem became distributing knowledge.
Then the problem became finding knowledge.
Now the problem may be connecting knowledge.
AI happens to be extraordinarily good at that.
The Library Can Finally Talk Back
Think about what a traditional library actually contains.
Thousands or millions of human ideas sit next to one another on shelves.
But they cannot interact.
A book about medicine cannot walk three aisles over and ask a book about engineering whether one of its ideas might solve a medical problem.
A military history book cannot compare battlefield logistics with Amazon’s distribution network.
A psychology textbook cannot spontaneously have a conversation with an economics journal.
All of those connections require a human being.
Someone has to read both things and recognize that an idea from one field might apply somewhere completely different.
Some of humanity’s greatest innovations have happened exactly that way.
AI dramatically increases our ability to make those connections.
For the first time, in a sense, the library can talk back.
And not only can it talk back, it can walk around the library pulling books off shelves faster than any librarian in history, which I suspect would make actual librarians either extremely excited or deeply annoyed.
Probably both.
But Didn’t AI Just Learn Everything From Humans?
One criticism of artificial intelligence is that it does not really create anything because everything it knows ultimately came from human beings.
There is some truth to that, but the argument becomes less convincing when we apply the same standard to ourselves.
Humans learn from humans too.
Every scientist learns from previous scientists. Every writer has read other writers. Every musician has heard other musicians. Every engineer begins with principles someone else discovered.
Einstein did not begin physics from scratch.
Steve Jobs did not invent the computer, graphical user interface, telephone, music player or touchscreen.
Shakespeare did not invent storytelling.
Human creativity has always involved absorbing existing knowledge, rearranging it, questioning it, combining ideas and occasionally discovering something nobody else noticed.
Isaac Newton famously wrote that if he had seen further, it was because he stood “on the shoulders of giants.”
That sentence could almost describe modern AI.
Except AI is standing on several million shoulders simultaneously, which admittedly makes for a strange mental image.
The interesting question is what happens when it begins finding connections that no human explicitly taught it.
If AI takes Idea A from one discipline, Idea B from another and Idea C from somewhere else and produces Idea D, which nobody previously articulated, is Idea D an artificial creation?
A human creation?
A collaboration?
We are entering territory where the language has not quite caught up with the technology.
There Are Still Things the Machine Has Never Lived
There is one enormous limitation to treating AI as the smartest thing on Earth.
AI has information about human experience.
Humans actually have the experience.
Those are not identical.
AI can read millions of pages about losing a parent. That does not mean it has buried one.
It can analyze leadership. It has never walked into a room knowing that the livelihoods of fifty families depend on the decision it is about to make.
It can study combat without feeling fear.
It can explain raising children without ever being awakened by one at three in the morning.
It can analyze love without falling in love, describe failure without lying awake at night replaying it, and study risk without having anything personally at stake.
Whether those distinctions will always matter is another philosophical argument entirely, but today they matter considerably.
Human experience gives information context.
Sometimes that context is irrational, emotional and messy.
It is also often where judgment comes from.
So Is AI Smarter Than Everyone?
This brings me back to the statement that started the whole thing.
“AI is smarter than everyone.”
I am not convinced.
AI is certainly better than any individual human at accessing an extraordinary breadth of information.
It can move between subjects faster than we can. It can compare enormous amounts of material. It can identify patterns. It does not get tired after reading three research papers and decide it would rather watch Netflix.
Those are substantial advantages.
But saying AI is smarter than humanity may misunderstand what we have actually created.
Maybe AI is not an intelligence sitting above human intelligence.
Maybe it is becoming an interface to collective human intelligence.
For thousands of years, humanity has been depositing knowledge into books, institutions, universities, databases and eventually the internet.
AI may be the first technology capable of walking through a significant portion of that accumulated knowledge and connecting it on demand.
That is not a small achievement.
It may actually be more profound than simply building a smarter machine.
Maybe the Future Isn’t Us Versus AI
Much of the discussion around artificial intelligence is framed as competition.
Humans versus machines.
Jobs versus automation.
Human intelligence versus artificial intelligence.
I increasingly think that framing misses the bigger opportunity.
Humans have qualities AI does not possess in the same way: lived experience, judgment, values, curiosity, responsibility, intuition and an understanding of consequences that comes from actually having something to lose.
AI has abilities humans cannot realistically duplicate: enormous information capacity, speed, cross-disciplinary pattern recognition and the ability to work through huge amounts of material almost instantly.
The interesting future comes from combining those things.
A capable person using artificial intelligence may ultimately be far more powerful than either the person or the AI operating alone.
And perhaps that is what we should actually be paying attention to.
Because maybe we did not build something that is smarter than everyone.
Maybe, for the first time in human history, we built something that allows everyone’s knowledge to begin working together.
If that is what artificial intelligence ultimately becomes, then the biggest question is not whether AI is smarter than us.
It is what happens when humanity finally has a practical way to access what humanity collectively knows.
Considering our track record, that thought is both incredibly exciting and just a little terrifying.
~ Ryan Gartrell
ryangartrell.com
