The Dunning-Kruger Effect Reveals why Most People Incorrectly Conclude They Can't be Replaced by AI
Mike Adams examines the Dunning-Kruger effect, where experts underestimate AI’s progress—like ByteDance’s upcoming video engine—while novices overconfidently claim irreplaceable skills. Studies show 94% of professors and 32% of software engineers rate themselves above average, yet AI already generates 35,000 books via BrightLearn.ai and outperforms humans in protein folding, art, and legal research. Corporations will adopt AI for cost-efficiency, even with imperfect results, forcing professionals to adapt or risk obsolescence as machines surpass human cognition in most fields. [Automatically generated summary]
This is a very important topic in the context of AI and job replacement.
I'm Mike Adams.
Welcome to this special report.
I'm an AI developer.
I'm a platform creator known as the Health Ranger.
And over the weekend, I spent a lot of time reading comments from people who are suddenly, suddenly realizing that AI is threatening to take their jobs.
And of course, I was trolling a lot of them too because that's always fun.
But no, I call it research.
I'm gathering information about their reactions.
So what I noticed is that, especially with Hollywood, because ByteDance has just announced their new engine that's about to come out, which is extraordinary.
This is an AI video engine that essentially makes Hollywood obsolete.
And it will empower individuals to be creators, to create amazing films all by themselves, without having to have a studio, without having to have actors.
They can create amazing films with original characters, and they can engage in compelling storytelling just by using AI video creation.
So what this Chinese company is doing, ByteDance, is they're empowering individuals with decentralized access to tools that allow them to become filmmakers, which is really remarkable.
There's going to be a lot of creative expression that comes from that.
But of course, the Hollywood types are losing their minds, and they were all over social media over the weekend insisting, insisting that they cannot be replaced.
That they are the only ones who know how to tell stories.
Oh, wait till somebody tells them that their ancestors passed down all human knowledge through nothing but storytelling.
You know, before the invention of writing and the alphabet, etc. It was all storytelling.
So no, storytelling isn't new.
It's been around forever.
And it turns out that it's not just Hollywood people that are good at storytelling.
As I said to one person, I said, look, it's false to believe that just because you're in Hollywood, that you are a special class of human who is imbued with the ability to tell narrative stories through the medium of film.
In fact, everyone can be a filmmaker.
Everyone.
Literally everyone can be a filmmaker.
And as you might imagine, that didn't go over very well with the Hollywood types who think that they have some kind of gift to that only they can create films and that they have the right to be the only ones who can create films.
Well, this is all part of the Dunning-Kruger effect and it applies to other professions as well, not just filmmakers and screenwriters, but also it applies to college professors and doctors and attorneys and coders and really just about any kind of profession.
The Dunning-Kruger effect is named that because of the psychologists David Dunning and Justin Kruger.
The Dunning-Kruger Effect00:08:35
And they worked at Cornell University and back in 1999 they ran a very interesting study testing undergraduate students on different kinds of skills like reasoning and grammar.
And they found that the students who scored in the bottom quartile, that is, you know, the lower 25%, when they were asked to estimate their own performance on the tests, they tended to estimate that they were around the 62nd percentile.
In fact, you know, substantially higher than where they actually scored, which turned out to be the, you know, on average, the 12th percentile.
And so this tendency for humans to vastly overestimate their own intelligence or their own skills or their own reasoning, this has become known as the Dunning-Kruger effect.
This is a very important term because it's going to become highly, highly relevant in the age of AI replacing humans in their jobs.
Why?
Because, of course, most humans, probably many of you listening to this, most humans believe that they cannot be replaced by AI because the things that they do are so high-end or so complex or so intelligent that they are irreplaceable.
And this is by and large the way that almost all humans feel.
This is their own perception.
But what the Dunning-Kruger effect actually tells us, and one of the key observations of the study, was what they called a dual burden.
It says that people who are incompetent in a certain area, they lack the ability to recognize their own incompetence.
And it turns out, thus, that humans who are not very intelligent are unaware of their lack of intelligence.
And humans who are bad at reasoning are not aware that they're bad at reasoning.
And humans who are horrible at whatever skill, right, law, they are unaware that they're bad at law.
And this isn't just in academia.
This study has been replicated at least dozens of times over the years, because that was in 1999.
There was another study that assessed gun safety knowledge at a gun range.
And it turns out that the participants there who knew the least about gun safety were the most likely to overestimate their knowledge in gun safety.
But that's just one study.
There are more, a lot more.
There's one that was in 1977 that gave rise to what's called the so-called better than average effect.
And it found, and this was in an academic environment, it found that 94% of college professors rated themselves as being above average compared to their peers.
94%.
Wow.
There's another study that found, this was studying drivers, you know, on the roads, people who drive vehicles, found that 93% of American drivers think that they are better than the average driver.
And then there was another study that was conducted among participants in a software company.
And it found that 32% of employees, almost one-third, believed that they performed better than 95% of their colleagues.
Wow, that's really astonishing.
Especially for those of us who are the smartest people in the room.
Of course, I had to throw in that joke.
Obvious satire there.
But again, if you're listening to this, you are among the smartest people, and obviously I am as well.
But how do we know that?
Well, it turns out that smart people are actually very good at gauging intelligence.
And shall I just say stupid people are horrible at it.
This is called the Downing effect.
And this is attributed to a researcher called C.L. Downing.
And he conducted a number of studies on perceived intelligence.
And he found that below-average IQ people vastly overestimate their IQ, while those people who have above-average IQ tend to underestimate theirs.
But he also showed that the ability to accurately gauge other people's intelligence was itself correlated with your own IQ.
So in other words, if you're a high IQ individual, you are better at determining the IQ of other people.
And if you're a low IQ person, you suck at that.
You will tend to be very inaccurate in appraising the intelligence of other people, either too high or too low.
And my own experience in this is interesting.
I'll tell you this.
If I had listened to a person for 20 seconds, I could tell you very closely their IQ.
You know, plus or minus 10 points, probably something like that.
At least that's my own assessment, right?
But you might say, well, no, I'm suffering from the Dunning-Kruger effect.
Okay.
Well, that effect is real.
But I'm confident that just by listening to the way people speak and the words they use, and also specifically their pronunciation, I can very quickly determine their intelligence.
And that's just something that I've picked up over the years, is pattern recognition in people.
And the funny thing about that is the way I speak is less intelligent sounding than what I actually am.
I actually tend to kind of, what's the best way to say this?
Well, because I strive to be approachable to every kind of person, and I'm not in academia, so I don't have the academia kind of speaking style, which makes you sound like data from Star Trek.
So I could have gone the full geek route.
Remember, I was a geek coming out of high school, and I could have gone to MIT if I had the money to go to MIT, but I'm glad I didn't.
But I could have been a really geeky speaker today, but instead, I'm who I am now, which doesn't sound geeky, mostly sounds a little bit peeved most of the time.
But nevertheless, that's just a self-observation.
What do you think about the way you speak?
Do you sound intelligent to other people?
And when you talk, do you try to talk in a way that makes you sound higher IQ?
Because that's a common thing, that many people do that.
Now, there's something else in all of this.
Back to Dunning and Kruger.
They found that the highest intelligence students, that is the top quartile, that they underestimated their own competence.
In other words, they didn't realize how bright they were because they assume that the tasks that were easy for them were also equally easy for other people.
And I'm guilty of this.
I've been accused of this.
It's like, oh, you say it's easy, but it's not that easy.
Okay.
I probably do assume that some things are easier than maybe they are.
But that's actually been captured by the Dunning-Kruger effect.
So I fall right into that.
Interesting, huh?
But there's also something called the imposter syndrome, which is where highly competent people fail to recognize their own talents.
They fail to recognize their talents.
You can look that up.
It's called the imposter syndrome.
Because it's like, well, somebody else must be doing these things.
It's very bizarre.
Now, here's what else is really interesting about this.
The Dunning-Kruger effect is strongly pronounced in America.
That effect, though, vanishes in Asian cultures, China, Japan, Korea, Taiwan, etc.
Because those cultures are more collectivism oriented, whereas in America, we are more rugged individualism type of people.
Dunning-Kruger in LA00:03:37
And so in America, and I've especially noticed this about Los Angeles and the film industry.
Anytime I've had interactions with people in the film industry, not everybody.
I know some great people who, like a friend of mine named Dan, he was a successful screenwriter at one point.
He's a sharp guy.
He's a realist.
But even he would tell you, and he lives, I think, around LA, but there's a lot of people in LA who they wildly overestimate their capabilities.
And they always are trying to sell a script or sell a movie concept or trying to pitch something, even though they're really not that talented.
For whatever reason, L.A. is, in my view, kind of the headquarters of the Dunning-Kruger effect.
LA has a lot of people who think they're smarter than they are.
And it has never been more apparent than watching film industry people talk about on social media how there's no way that AI could ever replace them.
Well, just you watch because I will replace you probably this year.
Because, I mean, remember, I've already built a book creation engine, BrightLearn.ai, published already 35,000 books.
I can guarantee you that I will be producing short, you know, mini-documentaries of video this year, probably before the summer.
That I know how to do the prompts.
I know how to create the scripts.
I know how to do all the recursive looping and fact-checking and how to use the APIs and write the code with AI agents and everything.
There's no question that people like me will make Hollywood obsolete very quickly.
Within, I mean, for full feature films, full-length films, that might be, it might be 12 months away.
It might be 18 months.
It might be 24 months.
I don't know.
Maybe on the outside, it's 30 months.
But it's not that far away.
People like myself and perhaps many of you listening, we will make Hollywood obsolete.
Because there's nothing actually that special about what Hollywood does.
It's just that they have a monopoly on it.
They have the money and the studios.
They have the unions and all the actors and everything there.
Well, guess what?
We don't need any of that.
We can create incredible movies doing just AI prompting.
And we're going to be featuring a lot of those films, by the way, on BrightVideos.com, which currently features all my podcasts.
But we're going to be very heavily, heavily focused on AI-generated film content at BrightVideos.com coming up as this begins to mature, which will happen this year.
So there's no question, there's no question that Hollywood is about to be replaced and there's a lot of resistance among the people who work in Hollywood.
And the way they sort of defend themselves against that is to think to themselves that they are smarter than everyone else.
No one else knows how to tell stories.
No one else knows how to make movies.
And they're about to find out that they're wildly wrong on that point.
And by the way, some of these studies on this, the Dunning-Kruger effect, have found that gender differences also impact this, that men in particular tend to overestimate their own intelligence relative to women.
So women are a little more accurate about their intelligence.
Machines Never Make Errors00:15:17
And women are also, I think, more accurate at estimating the intelligence of men.
Yeah, he's a dumbass.
You know, yeah, well, it should have been obvious, but now we know for sure.
So back to AI, I've actually seen people online, supposedly intelligent, mature people, saying things like, well, AI has never written a book.
And so I point them to brightlearn.ai.
Oh, well, here's 35,000 books.
Does that count?
Or they say, well, AI hasn't replaced anyone's job.
Oh, really?
Okay.
Why is Amazon laying off hundreds of thousands of people?
Why are all these companies laying off thousands?
Why is it incredibly difficult for any college graduate to get a job right now?
And why are coders being laid off by the tens of thousands everywhere?
Why is India's tech sector collapsing?
Because those are the coders.
Why is customer service collapsing in terms of human employment?
Because it's all being taken over by AI.
So I don't know what these people are thinking.
If they're saying that AI is not taking any jobs, AI hasn't produced anything.
They're living in a dream world.
Or they're living in 2024, let's say, when AI was just kind of barely getting off the ground.
Now, remember, I've been in AI for now almost two and a half years at this point, coming up on two and a half years, really intensely in this.
And almost everything that's happening now are things that I predicted on the record in this podcast.
You may recall, sometime last year I said that this year video tools would become available that would allow us to create three-minute mini documentaries in an automated fashion that were that were compelling, high-quality, lip-sync audio with persistent characters, etc.
And that's about to happen, it looks like, in a week, as Byte Dance releases their new system.
I also said that by 2027, we would be able to create full-length documentaries or feature films.
And it looks like that prediction might actually be too far out.
It looks likely that that's going to happen later this year at this point, but I guess we'll have to wait and see.
So I may have overestimated the amount of time it would take for that to mature, but I most definitely saw it coming.
And I said so in this podcast.
Now, one of the important signs of intelligence, of course, is your ability to model reality and to accurately project current events into the future.
Now, it's a dangerous art to try to do that because timing is always very difficult.
There are unknowns and there are factors beyond our control.
And I've been guilty of predicting many things too early.
You know, things that actually took many more years to come to fruition, right?
Although in this case, maybe my predictions are going to be too late.
Nevertheless, unintelligent people have a world model in their head that is not very accurate and thus they can really never anticipate what's coming next.
Highly intelligent people are able to anticipate.
People like you listening to this and myself and others, we're able to see what's coming.
Now, importantly, with AI, it is an exponential event that is taking place.
The rise in machine cognition is exponential and not linear.
And as you probably well know, the human brain is not built, the neurology just isn't there to accurately track exponential progressions.
Mostly, our brains think in a linear fashion.
Like, oh, if it was that way a year ago, then we just plot a straight line.
It's going to be this way a year from now.
The thing is, AI cognition is not moving on a straight line.
It's moving on an exponential curve.
So, oh, and by the way, the time scale of that is less than one year.
For example, AI inference costs were estimated to be, what, 40x cheaper each year, although I think that will plateau because of hardware scarcity, number one.
But also, AI cognition has been increasing at a rate that's more than, let's say, double per year.
It's more than double per year.
Or if you're familiar with so-called Moore's Law, which is attributed to a former Intel CEO that talked about the number of transistors that can effectively be applied to a microchip doubling about every 18 months.
And that law has also kind of hit the limit of physics, by the way.
But we are moving faster than Moore's law in terms of AI cognition.
So Moore's law said, again, number of transistors doubles every 18 months.
We are seeing a doubling of machine cognition much more quickly than every 18 months.
It might be, this is just an estimate, it might be every nine months, you know, half that time.
And that's extraordinary.
It might be less than that.
I don't know.
We're about to see some new models get released in the next week.
Not just the video model from ByteDance, but other major models.
And I believe we are going to witness a rise in machine cognition that is absolutely shocking.
I mean, psychically shocking to most humans because for the very first time, they will face, they will have to face the very real possibility that what they do in their job, in their career, in their lives, can be replaced by a machine.
That's a frightening thought for a lot of people.
And it's only now hitting people, even though I've been talking about this for two years.
For the most part, during those two years, like nobody believed any of it.
I would say, you know, machines are coming.
They're going to take your job.
People are like, ah, you're full of crap.
That's never going to happen.
Machines are never going to replace lawyers.
Now, machines pass the bar exam.
Oh, machines are never going to replace doctors.
Oh, you want to bet?
Doctors, humans are going to be obsolete.
It's already beginning to happen.
Oh, machines are never going to fill in the blank.
And now they're doing it.
It used to be a machine will never beat a grand master in chess.
Well, that happened more than a decade ago.
And then it was, oh, machines will never be able to write like a human.
Oh, guess what?
They do it all the time, every day.
Now, even on my website, it's brightlearn.ai.
And now it's like, oh, well, machines will never be able to do high-level math.
Yeah, they're solving major math problems now.
Some of the most difficult problems that have ever been created.
Oh, well, machines will never be able to do research science or chemical engineering.
You want to bet?
They do that in their sleep now.
Oh, machines will never be able to figure out protein folding problems.
Guess what?
Protein folding is one of the easiest things now that AI does.
On and on and on.
No matter what humans say machines can't do, oh, machines will never be able to create art.
Want to bet?
They're creating it in seconds now.
Well, they'll never be able to create music.
Oh, and then Suno came along, etc.
So everything that humans say machines can't do, AI ends up doing rapidly.
And people are just not keeping up with this.
They're not able to track what's happening.
So in summary, this all comes back to the Dunning-Kruger effect.
And we are now at a critical point in the history of civilization where for the very first time ever, and that's why this is different than every other invention.
This is different than the railroad boom.
This is different than the Industrial Revolution.
This is different than the personal computer revolution.
This is different than the internet.
This is different than dot-com.
This is different than every other technology that's ever existed in the world.
This technology is intelligent and it makes itself more intelligent.
This technology replaces human cognition.
And the only cognitive strategic defense against that is for humans to say or to think that they are smarter than they actually are.
Oh, machines could never do what I do.
It's so difficult.
It's like, what do you mean?
You work at a government office approving loans or grants.
You have a rubber stamp.
You sit around all day playing video games.
I mean, come on.
You don't even work.
I mean, that's like half the government workers out there.
I don't think they do anything.
How hard is that to replace?
I mean, seriously.
But the average person is like, there's never going to be able to replace me.
I'm an engineer.
I'm an architect.
I'm a doctor.
I'm a lawyer.
I have to have all this.
I went to law school.
I need all this knowledge of law.
Oh, guess what?
The AI model knows all law in all of history in every language in the world.
It knows every law of every state, of every country.
It knows the freaking laws of your homeowners association.
It knows the bylaws of your corporation.
It knows so much more than any human.
And it can operate on all of that in making a decision or writing its own, you know, whatever, its own bylaws or its own letter or its own case or whatever.
And I know people say, yeah, but they hallucinate.
They hallucinate.
Have you checked with humans lately?
Because humans are hallucinating all the time.
All the time.
That's the other thing I want to mention here, and then I'll wrap this up.
I keep hearing people say, aha, I found this one error in AI.
It's not doing it.
It's not doing it right.
I found an error.
Therefore, it will never replace humans.
And I say, have you looked at the error rate of humans lately?
Because humans are horrible with errors.
They'll make errors all day, all week long in their job, and they'll never even know they did because, Dunn and Kruger, they overestimate their intelligence.
They think, I did such a great job.
Actually, dude, you made 100 errors this week.
In the spreadsheet, too, in the accounting.
For God's sake, it's going to take us weeks to clean that up.
The bottom line, people are not as accurate as they think.
People are not as smart as they think.
This is almost universal.
And I've even suffered from some of the, I mean, I've listened to some of my own podcasts.
I'm like, I said that?
I used the wrong word.
Or I had the wrong number.
Like in my mind, I knew the right number, but when I was recording, I was distracted or I was looking at something on the screen or my dog was making a noise.
And I said the wrong thing, and I didn't catch it.
So I make mistakes too.
Humans make errors all the time.
So the question is not, is AI perfect?
The question is, does AI now make fewer errors than the people?
And also, cost differences.
For a corporation, is AI less expensive or more expensive than a human?
And the answer, of course, is way less expensive, fraction of the cost.
So it's not even a question like, can this AI do exactly what the human used to do?
No.
Even if it can only do half of what the human could do, that half can be accomplished at 1 100th the cost.
And then the other half, maybe it still has to be handled by a human, and that's fine.
But if AI can do half that job at 1% the cost, then guess what?
Corporation is going to bring in the AI.
They're going to do that.
So yeah, lots of people are going to lose their jobs.
And they're going to be shocked because they thought they were so brilliant and they're not.
They're really just not that smart.
But again, you know, most people think they're smarter than average or even smarter than 95%.
No, not most, but a third of the people roughly think they're smarter than 95% of the people.
And that can't be true, right?
Only something less than 5% of the people can actually be smarter than 95% of the people, right?
That's the only way the math works.
So there are probably like 28% of the people that are delusional, thinking they're the smartest people in the room, and they're not.
And those people are all over social media, too, posting about how they can't be replaced by AI.
Well, so the bottom line on all of this, folks, is I'm not here to put down humans, just to be clear.
I'm here to say that if you want to remain relevant in all of this, you need to upgrade your skills, upgrade your knowledge, and you need to be among the smartest and most capable people out there.
And most importantly, you need to know how to use AI because AI gives you plus 20 IQ points, in my view.
When I'm using AI, my IQ goes way up.
And I'm already starting out at a pretty good number.
It goes up 20 points compared to people who aren't using AI.
There's just no comparison.
You will be smarter when you use AI.
You will be more valuable in everything you do.
You will be more effective.
You'll be able to achieve what you set out to achieve in your life, in your corporation, in your nonprofit, in your church, in your mission, in your whatever, whatever it is you do.
You will be way more effective by using AI.
And sitting back and saying, I don't need to use AI because I'm already so smart that it can't help me.
That is not going to fly in 2026 and beyond.
Because it's just not true.
Nobody is so smart that they can't benefit from augmented cognition, which is what AI provides.
That's why the smartest people right now, the smartest people in everything, in art, in film, in music, the smartest people are all learning how to use AI.
I mean, I'm talking about the smartest podcasters, the smartest scientists, doctors, engineers, lawyers.
They're learning how to use AI.
And then they're applying it to what they're doing to either make their jobs easier, make their output more effective, catch their own errors, or magnify their expression, or other similar things.
So if you want to be more effective at what you do, learn how to use AI.
Otherwise, you'll find yourself very rapidly obsolete in this world, very rapidly.
Learn to Use AI00:02:55
It'll happen so fast, it will swamp you.
So, you know, where do you start?
Well, you can go online, use Replit, Replit.com, build something.
Sign up for, I don't know, use claude code.
I've coded almost everything with clawed code for months.
You can use, I guess you could use OpenAI, but I don't.
I don't like that company.
I don't use Google.
I use open source models.
You can download Quen, download various versions of all kinds of models out there.
You can run them locally on your own graphics cards.
You can download our model.
We released our model for free last year.
You can download that at brightanswers.ai.
You can run it locally on a graphics card.
It's brilliant.
It's got an incredible array of knowledge in it.
So check it all out and use our engines for free.
Again, brightanswers.ai, brightnews.ai for daily news trends and analysis.
Brightlearn.ai is our book engine.
And then we have the new site, brightvideos.com, which has all of my podcasts and videos.
Check that out.
You will learn a tremendous amount of information.
Bottom line, if you want to be well prepared for the future, definitely follow me.
I'm going to give it to you straight.
I'm in this arena every single day.
And I'll share with you what works.
I'll bring you the tools that work.
I'll help you stay up to speed.
You will never become irrelevant if you're tuned into my podcast.
You'll have the tools, the knowledge.
You'll know what's happening.
I'll tell you about the best engine.
I'm about ready to test DeepSeek version 4 when it releases soon.
I'm going to test the Byte Dance video model.
I did test Quin3 Coder next.
And I don't know, it's not exactly what I'm looking for.
It doesn't beat Opus 4.6 from Anthropic, but it's still pretty useful.
So yeah, I'll share with you what I know.
And as I learn and explore, I'll share that with you as well.
Help keep you informed.
Because look, we, all of us, you and I, no matter how smart we are, we are not going to be smarter than these machines.
I mean, seriously, you might be in the top 0.1%.
And I am too.
The thing is, it doesn't matter because the machines are going to be so much smarter than us.
They're already at that level in many ways.
And their capabilities will continue to increase dramatically.
And unless you can figure out some way to boost your IQ to 200, which I've never figured out how to do that, but even then, you know, the machine's going to end up at 500.
So it doesn't matter.
The key is to use machine cognition to augment your human mission, your human wisdom, your human guidance, your human inspiration, let's say.
Machine Cognition Augmentation00:00:24
That's what works.
I'll help you learn how to do that.
So stay tuned.
And you can also read my articles at naturalnews.com.
Thank you for listening.
Take care.
Astaxanthin is nature's ultimate antioxidant.
Experience the unmatched potency of one of nature's most powerful antioxidants with lab-verified astaxanthin supplements at the HealthRanger store.