Recently I gave two presentations in quick succession to the Dutch Network of Government Service Providers and the Dutch Advisory Council for Science, Technology and Innovation about AI. In these two different and brief presentations, I hoped to share some insights that would be useful to the people currently at the helm, the people who can or must make decisions about AI policy. Many thanks to the NPD and AWTI for the invitations and the stimulating discussions!
There is also
[a Dutch version of this post]. This article originated with these two presentations. The first, at the NPD, focused on the concrete challenges facing decision-makers. The second, at the AWTI, was about what is needed to build a trustworthy (European) AI infrastructure. And, for that matter, also touched on what that actually means. After these presentations I started writing, and eventually came up with this article.
While writing, I discovered just how challenging the subject really is. There are passionate advocates of AI, and there are people who believe it is the devil incarnate. These camps are so polarized that it is impossible to write something that everyone will be happy with. It is, however, entirely possible to write an article that makes everyone angry about some part. So here goes.
I do hope that both supporters and opponents will come away from this piece a little wiser. Perhaps there is even one thing we can all agree on: what exactly is the rush? Does every municipality really need to deploy Copilot this year? Did Dutch parliament really have to roll it out for everyone already?
What follows is not a “both sides” argument. For people who utterly dislike AI, you’ll find below that AI genuinely can do some remarkable things. For AI enthusiasts, I have a long list of reasons to think carefully about whether it is really such a good idea though.
Overall, I am deeply concerned that we are moving way too fast, and I don’t think that is an unreasonable position given the facts outlined below.
Spoilers — click here for an extended summary #
AI as a phenomenon is capable of truly remarkable things, and scientists won a Nobel Prize for creating a very useful one. Whether large language models are actually beneficial for organizations as a whole is another matter. Using AI to produce larger documents more quickly, only for other departments to summarize them again with AI, is not progress. AI is also highly disruptive to organizational career structures: where will your future senior staff come from if AI takes over junior-level work? And will senior staff really enjoy spending their careers checking AI-generated output?
Organizations are experiencing a strong Fear Of Missing Out that pushes them to adopt AI as quickly as possible. Yet despite how impressive AI really is, there is no urgency for most companies and institutions. If AI ultimately proves to be a good idea, it will still be there next year. Sometimes it seems as though people believe that if they don’t deploy now, they’ll never be able to catch up later.
AI comes with enormous problems and risks that deserve careful consideration. It emits absurd amounts of CO2 and consumes so much electricity that additional power plants are being built just to support it, all while we’re already struggling with severe grid congestion and increasingly destructive wildfires. The intellectual property status of the data that goes into AI systems is a massive issue, and it also creates uncertainty about who owns the AI’s output.
People brush these concerns aside surprisingly easily, but make no mistake: the AI industry will make them our problem. Try asking Microsoft whether they’ll indemnify you against intellectual property claims arising from Copilot’s output (a fun question to ask). Meanwhile, AI is disastrous for digital sovereignty: it’s surveillance capitalism all over again, imported from the United States. AI providers leak personal data on a massive scale, sometimes even data belonging to their own employees.
It’s difficult to know whom to listen to for advice. After investments totaling three or four trillion dollars, the AI industry itself is now so deeply committed that it can no longer be considered an impartial source. Much of the media, many consultants, and the broader software ecosystem have become so enthusiastic that it’s increasingly difficult to distinguish fact from hype. On the other hand, plenty of people feel so threatened by AI that they can only produce bad news about it. Then there are the moderates, who say, “AI is just a tool; it all depends on how you use it.” That’s true enough, but they never explain how to use it well, or if this is likely to happen, which makes the advice of limited practical value.
The pressure to roll out AI is so great that most organizations never define their expectations in advance or establish how they’ll measure whether an AI experiment is successful. As a result, they end up with many vaguely defined “successes” that may not actually be successes at all. From the executive suite, everything quickly looks impressive. But are your employees actually becoming more productive or happier? Nobody measures this.
There’s also the fundamental question of what we actually want from AI. It may seem appealing that AI makes writing reports and documents easier, but wasn’t writing those documents itself part of the thinking process that led to good plans? “Writing is thinking.” If you let AI do your writing, who actually learns anything from the exercise? Not you.
The companies providing AI services have also become financially precarious. There is considerable debate about whether the AI bubble is about to burst, or whether it is already beginning to do so. AI itself isn’t going away, but if today’s funding dries up, it could suddenly become much more expensive. That becomes a serious problem if your organization has, perhaps through staff turnover as well, become dependent on it.
In summary, the future of AI remains uncertain in many different ways. Its enormous potential (Nobel Prize!) is obvious, but that doesn’t mean everyone in an office should rush to have AI produce and read all of their documents. For most organizations, there is no train they’re in danger of missing.
A hasty decision to embrace AI, however, can cause enormous damage, to the climate, to your legal position, and to society through increased strain on the electricity grid. It can also drive away part of your workforce and leave you with an unsustainable organizational structure, because where will the next generation of senior staff come from?
In short: wise people don’t jump into the ditch just because everyone else is doing it. Take some time, wait and see, and you may spare your organization a great deal of disruption and unnecessary turmoil. A year from now, the picture will probably be much clearer.
AI: We Simply Don’t Know
The AI of today is not the AI of next year, and its societal impact is changing nonstop as well. It’s a genuine roller coaster. Still, there are ways to make sense of the confusion and arrive at a reasonable understanding of what is happening, and of what is, and is not, wise to do. This article hopes to contribute to that.
Anyone who claims to know exactly how things will turn out is, at the very least, confused. Take this PwC advertisement, for example:
Source - It says in Dutch “The value of AI is not distributed evenly. 74% of all AI value is captured by just 20% of companies”
Which reminds me that approximately 87.539% of all statistics appear to be made up.
A more honest graph comes from the Federal Reserve Bank of Dallas:
According to the “Dallas Fed,” AI will make the economy four times larger, or sixteen times smaller. The truth is probably somewhere in between!
So don’t take anyone who claims to know exactly what’s going to happen too seriously. That doesn’t mean, however, that it’s pointless to think carefully about how we should approach AI as a phenomenon.
The Nobel Prize in Chemistry #
Some people like to dismiss AI as nonsense. “Fancy autocomplete.” “A stochastic parrot.” Here’s a counterexample from biology, a field I happen to know quite well:
The creation of an AI model has, in fact, won a genuine Nobel Prize in Chemistry. And as Leiden University explains here, this was entirely deserved. Using an AI model, we can now predict the three-dimensional structure of proteins directly from DNA. This is something humans cannot do, and something computers had previously been unable to help us with in any meaningful way. That Nobel Prize was therefore more than justified. Biology and pharmaceutical research are already making grateful use of these new capabilities. The architecture of this AI model is remarkably similar to that of Large Language Models.
Some people argue that this AI is not truly intelligent, and “doesn’t know what it’s doing.” That’s largely true. But this AI application is already enabling discoveries that benefit all of us.
AI can also be remarkably good at hacking (although it doesn’t always hack what you intended). It can translate quite well and produce audio transcriptions that are better than human-level in many cases (although, once again, you really need to verify what it produces, and almost nobody ultimately does). You can also vibe your way into some very impressive demos.
So no, AI cannot simply be dismissed as “all nonsense.” But that is not a blanket endorsement either. In fact:
I Have to Mention It: The Lawsuits, the Casino, the Climate, Digital Sovereignty, and Grid Congestion #
People almost roll their eyes when you bring this up, but every major provider of Large Language Model services is currently embroiled in enormous lawsuits involving publishers, academics, media organizations, newspapers, and others. Under normal circumstances we would hesitate to do business with companies that apparently engage in such controversial practices. But when it comes to AI, that suddenly seems perfectly acceptable.
It’s genuinely amusing to ask AI vendors whether they’ll indemnify you against intellectual property claims arising from their output. Ask your legal department to look into it. The answer, or the lack of one, may surprise you.
By Catboy69 – own work, CC0 The financial condition of AI providers is spectacularly worrying. They’re all investing in one another in increasingly circular ways, and one of the largest players, Oracle, has already seen its credit rating fall to just one notch above “junk”. It’s quite something to build your future on top of a questionable casino. Debt issued by Elon Musk’s AI company is now also being traded as a “junk bond.”
The Financial Times wrote this week: “Whichever way, AI scepticism now feels like the consensus. As for what to do in preparation for a crash, you’re basically on your own. But if the music does stop playing, an awful lot of people are positioned and ready to say they told you so.”
We had also agreed, quite seriously, to reduce CO2 emissions. A great many countries signed up to that. Within a few years AI is expected to consume more electricity than France (or so it’s claimed), and old nuclear power plants are now being brought back online to satisfy demand. Apparently that’s fine too.
Closer to home, you may find that you can’t get an electricity connection for your new house in the Netherlands, because the grid is already overloaded. Yet we’re eager to build enormous AI data centers here, and preferably as quickly as possible. The implicit message seems to be: you can always live somewhere else.
We had also decided that we should work toward digital sovereignty and become less dependent on American Big Tech. Yet virtually every AI deployment in business and government amounts to American “surveillance capitalism,” in which our data is eagerly collected and exploited (and not for the purpose of improving our lives).
On top of that, most LLMs are built on the exploitation of underpaid low-wage workers. Equally fascinating is that AI advocates propose replacing your entire junior workforce with AI, without giving any thought to where future senior employees are supposed to come from.
And that’s all rather bizarre. Entire conferences and AI workshops are held where these glaring contradictions aren’t even mentioned.
Everyone currently beating the drum for AI is, in effect, also saying: “Too bad about your environment, your intellectual property, your digital sovereignty, your employees, and your inability to get electricity for your new home.” Because apparently we simply have to press on.
A rather curious state of affairs.
AI Can Also Do Astonishing Things #
If we momentarily set aside the climate damage, the electricity shortage, the intellectual property issues, and so on (as everyone else seems to do!), it turns out that many critics of AI have little or no firsthand experience with it. And that leads to fascinating clashes. Someone who couldn’t write a line of code before, but who spent the weekend vibe-coding an AI tool to organize their photo collection, comes back overflowing with enthusiasm. Then they run into a skeptic who insists that none of it is real and none of it works. Sparks fly. But AI really can give people an extraordinary experience. Someone who has never been able to program returns to the office on Monday with a fairly functional app. That’s genuinely astonishing. It would be like me swallowing three pills and suddenly speaking fluent French. I’d be wildly enthusiastic too.
Neo also suddenly knew Kung Fu in The Matrix
Anyone who’s had an experience like that comes away convinced: this is incredible! And for that individual, it genuinely is. I don’t want to downplay this, it’s remarkable what someone with no prior experience can now vibe into existence. You really have to see it to believe it.
At the same time, there is no guarantee that such an app will turn out to be maintainable, complete, useful, or secure upon closer inspection. In fact, it usually isn’t.
And that brings us to a central problem: a great deal of AI is being evaluated by people who simply aren’t qualified to evaluate it. Who can tell whether an app is actually good? That would be the end users, and the people who deploy and maintain these kinds of applications in practice. They know where the skeletons are buried. Do the backups work? Is the logging useful? Do users actually understand how to use the app? Can the help desk diagnose problems and resolve them? Does it crash on certain Samsung or Apple phones? Or, as happened with a recent Rabobank app update, does it get confused when you rotate your phone? Is the code readable and maintainable? Is the data storage GDPR-compliant? Where is the data actually stored?
It takes experience and expertise to judge whether an application is any good. AI enables many people to build things they could never have built before. Unfortunately, those same people then often decide for themselves that what they’ve built is good. That’s particularly problematic when the person in question occupies a senior leadership position.
The second challenge is that such a person often becomes convinced that the entire organization will benefit from their discovery. But has anyone actually asked the real users whether this is what they need? The organization may have entirely different problems than the absence of this particular app. Incidentally, even developers who don’t vibe-code often neglect to ask these questions.
The takeaway is that people, and especially executives, with personal positive AI experiences are not necessarily well positioned to decide on behalf of their organization whether AI has been a success. No matter how good it feels on a personal level. The same applies to people who use ChatGPT individually for work: it may feel great to have AI write and summarize reports for you, but is the organization actually becoming better as a result? That may not be your concern personally, but it certainly is the concern of management. I’ll return to that later.
An enormous problem is that in most organizations, people no longer evaluate whether AI is actually successful, or even can be. After all, it’s “the future.” AI is indeed here to stay, but it would still be nice to think carefully about what we should actually do with it. Or at the very least, define beforehand what you expect it to achieve and measure whether it actually does. Because “it felt good” simply isn’t enough. More on that later.
Nevertheless, AI Isn’t Going Away #
History is full of new technologies that people initially had no idea what to do with. Today the laser is indispensable, yet for decades it was famously described as “a solution looking for a problem.”
People also laughed at the Internet. As if anyone would ever buy shoes or pizza online “yeah, right.” But look where we are now.
But not everything becomes a success. We hardly hear about blockchain anymore, despite all the promises that it would solve every problem. Second Life, the Metaverse, and VR headsets also failed to live up to the hype.
AI, however, seems much more likely to resemble the laser than a cryptocurrency. That Nobel Prize also was no joke. And you really can chat with AI and ask it to write or edit documents for you. Whether that ultimately turns out to be AI’s most valuable application remains to be seen, though.
Having AI generate lengthy reports that are then read by other people using AI strikes me as a classic “horseless carriage.” The first motorized vehicles looked more like horse-drawn carriages without the horse than like modern automobiles. It took time for people to let go of old ways of thinking. I suspect the same is true of AI: right now we’re mostly using it to preserve existing practices (“documents that nobody reads anyway”).
My guess is that we genuinely have no idea what we’ll be doing with AI ten years from now. But it certainly won’t be nothing. And it almost certainly won’t be “horse-drawn carriage without the horse.”
Incidentally, even if the AI bubble bursts, AI itself won’t simply disappear. What will happen though is that it suddenly becomes much more expensive, temporarily relieving some of the pressure to use it. In fact, that increase in cost has already begun.
Who Should We Listen To? #
That’s not an easy question either. AI investments now amount to trillions of dollars. Entire industries, including the media, have been swept up in this tidal wave of money. For many consultancies and vendors, the success of AI has become existentially important.
They’re overflowing with enthusiasm, but then, they have to be. The financial interests involved have become so enormous that the AI industry itself can no longer serve as an objective source of information. It’s like listening to a nearly bankrupt used-car dealer enthusiastically pitching you their last remaining car. Of course they’re enthusiastic. And unless something very unexpected happens, many AI companies are going to go bankrupt. So the analogy is rather fitting.
Much of the media and many public commentators have also embraced the narrative that AI is simply inevitable. Entire “news sites” are now filled with AI-generated slop, so at the very least their shareholders clearly believe in it.
A Dutch government ministry recently gave a presentation stating that, for civil servants, “using AI is now mandatory” (provided it’s done thoughtfully). Mandatory, apparently! The Dutch governing coalition has likewise embraced AI wholeheartedly and without reservation, as we also saw in the coalition agreement.
Then there are, of course, the people who always seek nuance. They tend to offer compromises that aren’t particularly useful: AI is just a tool, and everything depends on how you use it. That’s true enough, for almost everything. But will people use it wisely? Do they know how? Are we teaching them? The moderates rarely discuss those questions. They also tend to argue that AI could work well under proper regulation, without mentioning that governments currently appear almost powerless in this area, and often don’t even seem interested in creating robust regulation, let alone enforcing it.
Something else we shouldn’t overlook is how today’s use of LLMs, for text, images, video, and source code, comes across to people who take genuine pride and pleasure in creating those things with their own minds. It’s not just that many of us think we can do it better than the AI. It’s also intensely irritating to have people constantly confront us with their latest LLM creations and expect us to respond. The subtext is obvious: “I can now do, with my buddy Claude, what used to be your job.” Your days are numbered! Perhaps they don’t mean it that way, but that’s certainly how it comes across.
It doesn’t help that we’re constantly told we have to get on board or risk being left behind. People who’ve never shipped an application in their lives tell me that my career as a software developer is over. The fact that software I wrote still powers half the Dutch Internet is dismissed as ancient history. I’m portrayed as some kind of tragic fossil who simply can’t keep up anymore.
When your life’s work is ridiculed like that, and your relevance is openly questioned, it’s extraordinarily difficult to maintain a balanced view of AI (though I do my best). And the entirely understandable resentment that results is then used by AI enthusiasts as an argument in itself: “You don’t have to listen to these bitter people.”
Painfully enough, that’s partly true as well. People who are, quite understandably, deeply worried about AI often produce an endless stream of bad news about AI, whether that news is accurate or not.
So if you want to form a well-founded opinion about AI today, you’re largely on your own. The industry, the media, the enthusiasts, and the critics all provide inconsistent signals that are difficult to make sense of. Even the nuanced folks can’t help you. Your own successful experiences with AI aren’t enough either, they cloud your judgment too much. And a personal dislike of AI doesn’t lead to balanced evaluations, either. In both directions, the same principle applies: feelings are not policy.
The situation is well captured by this quotation, by the way:
“In
[his essay on Salvador Dali], Orwell argued that because Dali was a repulsive human being, the right wouldn’t admit that he was a great artist; conversely, because he was a great artist, the left wouldn’t admit he was a repulsive human being. I’m seeing the same thing with AI: because it’s unethical, one side won’t acknowledge that it’s useful, but because it’s useful, the other side won’t acknowledge that it’s unethical.” –[Greg Wilson]
AI Experiments in the Workplace and in Education
The pressure is enormous, so everyone is rolling out AI in one form or another. Because apparently it simply has to happen, even when there’s no plan (read this piece!).
If you’re running an AI experiment, make sure you actually measure whether it’s a success. To do that, you first need to define what you’re trying to achieve, and how you’ll measure whether you’ve achieved it. The enthusiasm is so great that many AI experiments begin without any plan at all. The result is conclusions such as “the summaries look good,” or “24% of employees participated in the pilot.” Or sometimes not even that.
Better metrics would be things like: “75% of the minister’s commitments were captured in the AI-generated summary,” or “25% of callers reported an error in the transcript of their conversation.”
Given the enormous expectations and pressure surrounding AI, proper evaluation is not something you can simply do as a hobby. There are specialists in workplace productivity and organizational effectiveness, and you should involve them before deciding, based purely on vibes, that your AI experiment is the an astounding success.
It’s remarkable how many AI pilots begin without anyone first deciding what the objective actually is. The pilot may then conclude with a “pointless cake celebration,” without anyone knowing what, if anything, has actually been achieved. It may even have made things worse.
As an example, here are the results of a study into Dutch municipal chatbots that are in production:
Source - 64% of questions were answered incorrectly, 10% correctly This result could easily have been avoided.
Another problem is that management often announces an AI pilot with great enthusiasm, making it clear that they expect big things from it. Afterwards, employees are enthusiastic too. Because they saw senior management setting the example!
If you genuinely want to understand what’s happening on the work floor, which has always been difficult, by the way, you’ll need to conduct a credible anonymous survey. And you’ll also need to think about second-order effects. So don’t begin rolling out AI until you’ve defined what success looks like, what you expect from it, and how you’ll measure whether those expectations have been met. Otherwise, it’s all too easy for every experiment to become a “success.”
What Do We Actually Want?
Apparently, AI! And as quickly as possible too!
Still, it’s worth pausing to ask what an organization is actually trying to accomplish. A chatbot is not an objective in itself. Helping citizens or customers more quickly, more cheaply, and better, while remaining accurate, is an actual worthwhile objective.
It’s certainly possible to use AI to produce enormous reports and documents at record speed. Then other people in the organization can use AI to summarize that stream of documents so they can understand it. That is not progress, even if no information were lost or hallucinated in the process, which, in reality, it often is.
People have long observed that writing and productive thinking are intimately connected. “Writing has been called the process by which you find out you don’t know what you are talking about.” Likewise, “Writing is thinking.” Another favorite: “If you’re thinking without writing, you only think you’re thinking.”
Many internal documents ultimately aren’t read a lot or very well. But they are presented, discussed, and sometimes implemented. Until now, writing those documents has been an integral part of an organization’s thinking process. Formulating them made both the presentations and the plans themselves better. Even if a document wasn’t widely read, the fact that someone had written a substantial piece about an idea was a strong indication that they had genuinely invested time and thought into it.
With AI, however, it’s possible to produce superficially convincing documents with very little effort or reflection. Such a document is no longer evidence that someone has thought things through or put in significant work. And perhaps the results aren’t actually that good, either.
My impression is that many organizations are unlikely to improve themselves if less genuine thinking takes place, or if the informal “did someone actually put real effort into this?” test no longer works.
That’s before we even consider the damage done when a public policy document turns out to be full of hallucinated references that don’t actually exist, something that even consultancies writing about AI keep doing in their own AI reports (additional link). Because we must definitely also not forget that even the best AI frequently generates absolute nonsense.
It’s therefore worth thinking very carefully about what you actually want AI to achieve before you begin deploying it.
I recently attended a presentation by people who were using AI to read millions of newly written contracts so that the transactions described in them could be entered into a database. That’s a rather strange use of AI in 2026. Why aren’t those transactions simply entered into the database digitally in the first place, instead of being inferred from contracts written by humans? Or perhaps even contracts that were themselves drafted by AI. It could all have been done directly.
At the same time, if you have millions of historical contracts whose contents aren’t in a database, it’s hardly “better” to have people read them manually and copy the data over. That’s not enjoyable work either. Verifying that the AI extracted everything correctly is difficult enough.
Another nice example I heard involved a government agency that wanted to use AI to automate part of a process, but quickly discovered that the input data was far too messy for this to work. Once they cleaned up the data, the existing process improved so much that they no longer needed an AI version at all. Of course, that won’t always happen, but it’s worth remembering that AI isn’t magic: it requires good input. Many companies discover this when they try to build a chatbot that understands their business processes. If those processes aren’t documented properly anywhere, the chatbot won’t perform miracles to fix that. That’s when hallucinations become a real risk.
Doing “AI for AI’s sake” is an expensive hobby, and a risky one. Plenty can go wrong. It’s perfectly reasonable to spend some time thinking about that first.
What Is the Right Pace?
Right now, organizations are under enormous pressure to “get started with AI.” After all, who wouldn’t want to be seen as a modern organization?
Still, it’s worth taking a step back and evaluating all this urgency. First, the downsides of moving slowly:
- You may miss out on potential cost savings.
- Some employees may simply expect AI to be available.
There is also a general fear of “missing the boat.” But many organizations (governments, for example) are not locked in a fierce battle with competitors. That boat will still be there next year. AI is not going to disappear overnight. And so far, organizations do not appear to be seeing measurable productivity gains from AI. That may well change, but for now you’re not missing much.
Point 2 is a more significant concern. Someone who has spent years creating and analyzing documents with AI may no longer be able to do that work manually. This surprised me too, but someone from a large HR department told me about the problem. A new class of employees has emerged that no longer functions effectively without AI assistance. Their own skills deteriorate rapidly through cognitive off.
Moving too quickly has downsides as well:
- Violating climate agreements, the questionable intellectual property situation, and broader ethical concerns.
- Employees may leave because of AI.
- Declining quality of service.
- Violating the privacy and rights of citizens, customers, or other contacts.
- Further erosion of digital sovereignty.
- If the AI bubble bursts, everything may suddenly become unaffordable.
- Reduced organizational thinking capacity.
- Disruption of the career pipeline (“all the juniors are gone, where will future seniors come from?”)
Regarding point 1: from the executive suite, things often look very different than they do on the work floor, where employees hear their director enthusiastically announcing that their jobs will soon be automated. That does not necessarily end well.
Regarding point 2: see the chatbot examples above. Things often do not improve, even if they are now considered “modern.”
Regarding point 3: AI experiments have repeatedly been shown to leak enormous amounts of personal data. Meta (Facebook), for example, recently leaked sensitive employee data during an internal AI trial.
As for the rights of your customers, applicants, and other contacts, it’s easy to forget that we have both the AI Act and the GDPR here in Europe. These laws impose specific rules on how AI may be used to screen and reject job applicants. Before you know it, someone in HR is using AI in ways you never approved, yet your organization is still in violation. Despite what many people claim, major provisions of the AI Act really do come into force this year.
Regarding point 4: we are supposedly trying to reduce our dependence on American technology companies here in Europe. Yet the overwhelming majority of workplace AI comes from Big Tech, and from companies that are eager to collect, store, and reuse as much of our data as possible for future AI training.
I discussed the reduction in organizational thinking capacity above. Simply writing down plans is itself a powerful way to think more clearly. If you let AI do the writing, you lose that valuable benefit.
As for the final point, that brings us to the next section.
The Career Pipeline
If we let AI take over junior-level work, we have to ask where future senior employees will come from. Honestly, I have no idea. But if your organization relies on a structured career ladder, it would be wise to have an answer. Damage to your talent pipeline takes years to repair. The promise is that AI can perform much of the work while humans simply supervise it. But that fundamentally changes people’s jobs, and it turns out humans are not particularly well suited to spending most of their working lives “checking what the computer did.” You may end up driving away your experienced senior staff, who have little interest in such a role. Meanwhile, the juniors are no longer needed because AI has replaced their work.
This may sound dramatic, but it is, in fact, what is already happening.
When organizations declare, as the Dutch cabinet has, that they are “fully committed to AI”, they also need to think through the broader consequences. Will we still have a healthy, functioning organization a few years from now?
“It Always Works Out!”
Whenever you raise concerns about where future senior employees will come from, AI enthusiasts almost always respond that these things always sort themselves out. New jobs will emerge, jobs that didn’t exist before. “The market will find a way.”
They often invoke the Industrial Revolution, but neglect to mention that it took multiple generations of hardship before things “worked out,” and in the meantime conditions were dreadful.
Others make unproven claims that they now have more time to think because AI handles the writing. Still others argue that AI may not generate genuinely new ideas today, but surely it will in the future, and therefore it’s already acceptable to let AI do the thinking now.
These are attractive claims, but they also require a fair amount of faith to accept. Show me the evidence please.
There is nothing unreasonable about thinking carefully before reshaping your organization around AI, even if people incorrectly mention the Industrial Revolution as proof that everything always turns out fine in the end.
So, Now What?
As of July 2026, one thing is fairly clear: the hype surrounding AI is enormous, but where it is ultimately headed remains highly uncertain. There are also compelling legal, ethical, environmental, and financial reasons to approach it with caution. On top of that, AI may have serious and unintended consequences for both the workforce and the privacy of citizens or customers.
Criticism like this is often dismissed with, “Well, that’s just someone who hates AI.” But that’s far too simplistic. We’re talking about what may become the biggest transformation of the workplace in history, and organizations are rushing into it at remarkable speed. There are also plenty of objective reasons for concern (see above). Those deserve careful consideration, which means taking criticism seriously rather than dismissing the messenger.
The good news is that, for most organizations, there is no real urgency. If you don’t roll out Copilot in 2026, you’ll still have the opportunity to do so in 2027. By then, many of today’s uncertainties will have become much clearer. The major lawsuits may have progressed, several of the weaker AI companies may have gone bankrupt, prices will likely be higher but also more realistic, and the practical impact of the EU AI Act will be much better understood, which may save you from costly compliance mistakes.
Perhaps even more importantly, by then we’ll have learned a great deal from the organizations that rushed in first. Did they actually benefit? Or did they instead end up running into serious problems?
In the meantime, if your organization decides to experiment with AI, don’t make the same mistake so many others are making today. Define beforehand what you hope to achieve with the AI experiment, and specify exactly how you will measure whether it was successful. “It felt pretty good” is a rather thin justification for fundamentally changing how an organization operates. If you evaluate your experiments rigorously, you are much less likely to regret them later if they turn out not to have been worthwhile.
And once again, don’t forget the fundamental issues surrounding AI and the AI companies themselves and the way many of them operate. When you adopt AI, you are also choosing to do business with those places. The least you can do is acknowledge that fact (as I did in translating this piece).
Good luck with it. AI is not going away, but how this story ends is still, to a significant extent, up to us.
P.S. Feedback on this article, positive or negative, is very welcome. I may well have overlooked things. You can reach me at bert@hubertnet.nl
Before anyone asks, this article was mostly translated into English by ChatGPT, after which I edited it thoroughly to fix the nuances the AI got wrong. I could write a whole separate article on why I chose to do this, but for now I hope you can focus on the content. If you want to rationalise it, lots of people were getting
[the original Dutch post]translated individually, which surely used even more energy.
Further Reading
AI Mania Is Eviscerating Global Decision-MakingThe AI Collapse Pre-Mortem- My own introduction to deep learning: Hello Deep Learning