The Better AI Gets, The Worse You Become. Here's What To Do About It.

One of the things that, uh, fascinated me with your work is the idea of the better.

A automated system is kind of the worse a human gets. Can, can we start there? Jump straight in. Can you tell us what you found there?

the more we automate the, the less the human is able to support that automation. Um, we can look at multiple disasters globally, um, from Uber's self-driving vehicle accident, or the, uh, operators relegated to, to a couple different input tasks, but not really capable of attending to the road, to,

Hmm.

in a time where we're looking from an enterprise perspective into automating a bunch of different processes, right?

Through ai. Um, your background is like, you come with quite a pedigree. Maybe it's, it's a good idea to spend like two minutes on, on what have you done in terms of academic studies to, to be able to, to kind of state the predictions and the, and the, uh, different opinions that you have here today.

So I've been working in areas of transportation safety, leveraging AI tools for years, working and, and strategically using and, and thinking forward in what AI may be able to solve for over 25 years.

and I think. From reading your, might be good to mention as well, you just released a, a new book on, on AI and then, we'll, we'll get into that soon.

That's a great question and, and I think that as humans we are much better predicting the future. Then we are protecting the timelines. You know, there are a couple of illustrations out there of, of technologies that seem to just work overnight. I mean, the iPhone is, is the one that sticks in, in, in many of our minds, but these are rare occurrences.

If we assume that the learnings from, from the self-driving cars and how the human skills kind of atrophy over time, what do you see as the enterprise version of, of a driver who's forgotten how to drive a car?

that's a really good observation and pickup here, Johan, and I think that is very well known that skill atrophy is going to occur. Um, you can just look at the studies, um, in, in recent months in neural activity when you're using, uh, chat bot. But I think that where studies do fail is that it's, it's it's relegating to the chat bot.

you could say that the car, in the non self-driving version of it, it's, um, decreased our ability to run long distances probably.

You know, that's a really good question.

Hmm.

on the other hand, many individuals, um, suffer from a, a, a, a difficulty being able to ideate, come up with a new idea and think about an approach.

if you're a CEO. So what I hear you saying is fundamentally like either you can focus on raising the floor, the, the worst performers in turn in the analogy of the spellchecker, right?

I think it's both. Um, I think some individuals are, are going to think about new ways and accelerate your organization forward.

Yeah. And it's interesting, like there's a certain sample ignorance when, when you start upgrading. Your text for the first time with chat d pt and you ask it to, to refine arguments and, and this is the target audience and so forth. You see a massive improvement typically because we're all not authors that are brilliant at writing.

But look, the same is true for your podcast here, is that you're bringing on different guests to change the tone of the situation.

Yeah, it does.

So you, so, you know, that's not success. Success is bringing human creativity into this. And, and, and that's where the, the inter fusion of, you know, human skill and, and machine expertise can really occur together.

I I think you make a good point in terms of work slop and, and taking accountability completely outta the room when delegating to ai. And I think there's a conversation that needs to be had there. We also have the other side, like the secret cyborgs. So people that have automated stuff, but they don't wanna tell anyone because they know that the only reward that they'll get is more of the same work.

So my message to a few, few senior leaders at this point is that, look, we need to embrace in, in our organizations. To think ahead. Um, traditional academic structure is not gonna teach people how to use these tools.

Moving back to the, uh, to the CEO and, and what's on top of their mind. I, I think everybody's seeing this massive change on the horizon, but we see very different responses.

the search engine, you know, concept that, that we had many moons ago, um, in AltaVista was, we're gonna ask this anything, it's gonna produce an answer. And, and I think that's really a, a, a great parallel to modern chatbots. Um, here years ago we had the idea of, of what, you know, AI and search could provide, but it took us decades to really embody that. and I think it's a great illustration of, we have some, some ideas of where the future, whether that's driving, whether that's information technologies, whether that's augmenting our move away from the smartphone to some other parts, smart personal device, you know, we have visions, but the embodiment and execution of that.

Having said that, it's interesting on, on the kind of blind tests, especially on the ability to show empathy. Uh, and I know there's been studies done in, in the medical field that that AI interactions score higher on the empathy test. And, and surprisingly to me, I think it was in McKinsey's latest AI at the workplace, one of the biggest use cases for AI is like personal partnership and therapy, which it was completely unexpected to me.

That doesn't surprise me because many of us fear there's things we want to talk to a human about and there's things we're embarrassed to talk about to a human about. Yeah. So you, we've been asking Dr. Google everything for years, right? Mm-hmm. Why? Because I don't wanna call a clinician and ask about stuff that I'm not so sure I wanna describe this about myself.

Yeah.

So, you know, if we look at automated driving systems, whether we talk about cruise, um, whether we talk about Uber, um, you know, no one who's out there, has sustained their, um, moment in the negative sun. Um, when, when something occurs, every organization that is. Know, perforated robots and, and safety centric environments like driving has disappeared.

Yeah. And I think the handling the exceptions is also a really interesting point. And similar to, to your 1% when, when automation fails from the automotive industry. So when it comes to, Making available, like light level of therapy through chatbots.

And the same's true in on the clinician side, the, the human is not always right.

So I, I read, I think it was Neil Degrass Tyson who had an example around. So, uh, traffic fatalities with deer accidents in the US is quite high.

Yeah. And that I think will change over decades. So I think we will begin to trust robots. I mean, the Japanese society in particular has trusted robotics much more than a lot in the rest of the world.

Mm-hmm. Interesting. Yeah. And, and

I think that will occur over time. Um, so I think the concept of therapy robots has been in Japan for, for years. You know, that'll occur around the rest of the world. It's just gonna take time. We evolve and machines tend to change quickly, but humans evolve over time. And, and for some of us, hey, we're lead adopters, we'll move a little faster.

What are some really. Practical, you need to unlearn this type of advice, especially to leaders?

it wouldn't surprise me, 75% of major organizations out there still aren't using AI in ways that make sense. Um, I had a phone call from, from a friend the other day, you know, a multi-billion dollar company and they have not touched AI and they don't allow their employees to touch ai.

Have you written or thought a lot around, uh, kind of how we measure productivity for knowledge workers?

yes, and I think it's one of the key pieces in how to make AI useful.

Hmm.

Okay. You know, remembering that most failures in AI are not really technical. The AI does what it was programmed to do. They're human. Hmm. So, you know, the systems have to be well thought out before you can automate 'em. You know, and this is why automated driving has been, you know, a few years out for, for 20 years now or more.

earlier on we talked about like what, what are the, um, what are the, the human skills to double down on, um.

That's a really good question. Um, and I don't think it's a one size fits all answer.

Yeah, fair.

Some of us,

Yeah. Now I was thinking back to my own path and, and the, the years that has gone, because I think I've been leveraging down things that I already was quite strong at. 'cause those things are, are also easier for, from a motivational standpoint, I get better at what I'm already good at.

the, the word you used partnership is really critical here.

Hmm.

Um, but as opposed to AI augmented or ai, um, you know, uh, co-pilot or co-creative. Um, so I think we need to be thinking about whether it is a pure regression to the mean, based upon AI's ability to interpret everything that was given or. Or or something that came from, from human creativity.

One of the things, uh, being very practical here that, that I got a lot of value out of was spending a lot of time kind of categorizing together with ai what are the patterns of my thinking? So whenever I approach a new question, how do I approach it?

I'm a ChatGPT user and you know, the, the memory of what I've done in the past, have I ever written something like this? I can't remember what I wrote a year and a half ago. Yeah. Um, you know, I can go searching for hours on that and yes, you have no you haven't.

So I, I freeze and I think probably that's true for, for quite a lot of people, sadly. again, coming back to the, the kind of senior leader audience. So there's a balance between rushing ahead, implementing things without proper forethought into, um, safety standard or security practices or whatever it might be.

Uh, that's, that's an important point. I think there are a lot of organizations that are far too in on the, the AI financial bubble of today. Um, I think, you know, they've gone way too far over the edge in, in their investments and they ROI required on those investments just to, you know, it, it's quite frankly not going to occur.

it's like the actual investment in ChatGPT, for example. The

ChatGPT is just one of those. I think there are many other organizations out there that, that spent just fortunes and, and, and, and when you think about the math of the ROI, it would have to be so massive that it, and what's important about that math is that, you know, it's also, it's kind of assuming the technology's not gonna evolve from here fast enough.

if we were to summarize 2025 from the AI and in the enterprise space, what was the completely expected outcomes and what was the unexpected outcomes?

So the completely expected outcome to me is that this, this balloon keeps being pumped, uh, solar and fuller of, um, hot air and overinvestment. Um, but I think to me, there's this shift. The unexpected is there's an embracement and shift in the populace that its use of AI tools occurring. People are seeing value in what they can do and what they can help with.

Hmm.

I think 2026 becomes the year we focus on the human's role and, and really human-centric ai, you know, after years of, so over-focusing on the technological advancements. I think the new frontier in, in, in 26 becomes the human factor here. Yeah, okay. We can get better. Um, you know, so I think organizations that are leading here will recognize the competitive advantage doesn't come from deploying more a AI to replace humans.

I think that's a super interesting point in itself. So the more familiar and almost to the point of experts people get with these tools, I see two major shifts. One One is, as a consequence of the technology, you start reflecting way more about the essence of humanity.

Yeah. Well, under this, you know, move towards a human-centered framework and, and in and I, and I think 2026, I think it's the year it's human again, if that makes sense. Um, and I think that you're gonna look for a better understanding of what AI should do for us. Hmm. Uh, I think that the, the fundamental question is moving beyond the fascination of, of technology to, to thinking about is AI generally useful and where, and, and, and that really revolves around, you know, a lot of the thought points.

because it's also a question of, of how do we integrate with this technology in a not just useful way, but actually healthy way.

Yeah. And, and you look at the cover Time magazine, you know, the person of the year is ai. Um, you know, and, and I, but I think the person of the year in 2026 is 180 degrees from that. It's, it's a human, um, you know, the role of, of a human.

it's interesting you talked about the a 35 hour work week before, from a individual's perspective, that'd be awesome. Right. But I wonder how the politicians of the world look at this because it's kind of dangerous as well to, to, if we're in a, in a global economy where we need to compete with, with everyone, right.

Look, the realistic side is that you, you have lived in Sweden for some time, while the weather is far from perfect during nine months of the year. It's gorgeous during the summer. Um, the reality is the Sweden's take enormous part of that summer off and swish summer is, is four to six weeks, but they come back retuned to think more effectively.

Yeah. I think that's a really good point. And it's easy to kind of, um, take that into, to your own life, right? Just work more isn't necessarily the best answer to, to get the best output of your life. Like periods of taking a step back, periods of reflection. I think a lot of high performers the last five years has discovered meditation, for example, as a brilliant example, how to get to, to the kinda essence and the clarity of, of what needs doing, not just we need to do stuff right.

very rarely will you find an individual who says, I want wish I worked more.

Yeah. You have that expect as well. You come back and

say, okay, the expertise says, I wish I lived more. we are only on this planet for a short period of time.

We want to ensure that we're not just profitable this quarter, but that we can take care of, of our employees and, and a lot of these more value driven companies might, if you play this out. I'm being a little bit optimistic here, I think, but this is how I want to think, at least if you play this out, it's more about understanding what can AI do for us as a company over time, rather than just to technology.

I think that's possible. I think there's, there's, there's, there's regulatory efforts in Europe that, that are problematic as well.

And it's interesting when, when we speak about the AI bubble there, there's a lot of talk around the AI bubble right now.

I, I firmly agree. Look, I think the technology is a balloon. It's gonna go up and down. I think, you know, we'll find new uses for the technology.

Hmm.

You know, you can bet that, again, you know, we're, we're, we're having the wrong conversation. But I don't think anybody knows, I think the investments are being made kind of blindly. Um, we have to keep up with the Joneses being the model. Yeah. As opposed to, okay. We do have a, a long-term value proposition for our shareholders here.

We will find ROI it just may have nothing to do with the companies that are investing in the infrastructure today. Yeah. That's interesting. And, and, uh, talking about the open AI as an example, like the business model is, is kind of. Shaky, right? Because you have your, your core product gets, I don't know how much better each year, 40 times better per year or something like that in, in raw compute. But the cost goes down like eight, nine times like that.

we don't know inside Sam Altman's mind what their ultimate goal is. I mean, look, is the ultimate goal of open AI to, to, to provide a new search engine that transforms the internet economy and economics away from, from Google.

Yeah. That's such a scary proposition. I wasn't actually aware of that until they did the code red that they pulled back the, the kind of a push towards an ad model in, in ChatGPT.

most of us don't realize how much negotiation is going on in the back room for our attention. I mean, you, you click and search for something is that there's a negotiation going on, uh, for fractions of, of cents on what ads to put in front of us.

outside of, of the year of the human, what are some predictions for 26? It's been a lot of talk around the agentic AI during 25, and then once we've starting to try to industrialize and scale the, the kind of proof of concepts agent has been quite difficult.

Look, agen is just automation by a different name to me. I mean, it, it's cool name, but we're trying to to to automate a lot there. And, and, and, and, and I think that we are moving too far on the automation side and a lot of folks with use of Ag Agent ai.

Do you see, I know this is probably not your area of expertise, but, uh, we haven't seen that much movement in the, the kinda legal area.

I don't think you're gonna see the legal sides.

Yeah. And there's a lot of human bias. And it comes kind of back to, to your idea of we're more fine with humans making mistakes than we are robots, uh, from from our conversation.

Yeah. Look, look, we, we look at traffic enforcement one, you know, we worry about the bias of who the law enforcement is gonna pull over, but we're even more concerned with the bias of how the machine intelligence may take it. Mm, absolutely. There's a traffic light, there's a traffic camera, and I believe in the US the other day, that issued a couple thousand tickets because the speed limit number in the traffic camera was wrong.

Yeah. Yeah, exactly.

because that's because the two of us are examples of individuals who, who are moving and leveraging these tools as co-pilots in life to advance what we're capable of doing on a daily basis.

A hundred percent.

We're not relegating to the robot, we're using the robot as a collaborator. Now, quite frankly, I'll, I'll frame it to you this way, without ChatGPT, you would've had a TE team of 20 and needed funding levels to create what you're doing now.

Hundred percent.

That, that now. You're able to automate and do faster.

That's fantastic. Hey, to close, close this, uh, this conversation out. What is the one most important advice that you'd give to, to like the individual? Like what are, what, what's the New Year's resolution to get the most out of AI for 2026 to make?

I think that the, the New Year's resolution is learn to play more.

Hmm. Fantastic. Brian, thank you so much for coming on. This was fascinating to take part of a lot of nuggets of, of wisdom and I, I really do hope that 2026 becomes the year of the human.

I, I, I think that would be the best thing for society in general is we think about us, um, as opposed to it's all about technology. At the end of the day, let's use technology to help us as a human species and a human race evolve. And, and that's why we started inventing things.

Thank you so much. That's a, a good note to end on for the theme, how to make AI useful. Thank you, Brian. Thank you.