#541 The Leadership Echo Chamber: Why Artificial Agreement Is Becoming a Strategic Risk | Article by Niels Brabandt
The Leadership Echo Chamber: Why Artificial Agreement Is Becoming a Strategic Risk
Article by Niels Brabandt
Artificial intelligence has moved from novelty to infrastructure. It is now part of daily work, executive preparation, strategic analysis, market evaluation, talent management, communication, and decision support. For leaders, this is both a significant advantage and a serious risk. AI can sharpen a decision, widen a perspective, challenge weak assumptions, and help organisations act faster. Used badly, it can do the opposite. It can make weak decisions appear stronger, confirm biases more elegantly, and turn an existing leadership echo chamber into a technologically enhanced one.
This is the core risk explored by Niels Brabandt EMBA MBA MSc in this week’s leadership podcast and videocast: the leadership echo chamber and the rise of artificial agreement. The term describes a situation in which leaders do not use AI to challenge their thinking, but to confirm it. They enter the conversation with a preferred answer, ask a shallow question, receive a polished confirmation, and then mistake fluency for validity. The result is not better leadership. It is a more sophisticated form of self-confirmation.
AI is not the problem. Untrained leadership use is the problem.
The starting point must be clear: using AI is not a sign of weakness. On the contrary, modern leaders should use the best tools available. In a serious organisation, it is reasonable to expect people to use AI to test ideas, accelerate preparation, organise information, improve drafts, and challenge their own assumptions. Pretending not to use AI may soon look less like professional integrity and more like technological denial.
The problem begins when leaders use AI without competence, structure, or intellectual discipline. A one-line prompt, written from a biased position, is not a decision process. Copying and pasting a fluent answer is not strategic judgement. A chatbot response is not a governance framework. The availability of AI does not remove the responsibility of leadership. It increases it.
Niels Brabandt argues that the real danger is not that AI is available. The danger is that leaders may use it badly and still believe they have improved the decision. This is where artificial agreement becomes dangerous. The leader is no longer surrounded only by human agreement. They may now be surrounded by algorithmic agreement as well.
The 24/7 advantage creates a behavioural trap
AI has one unbeatable organisational advantage: availability. It is there at 3 a.m. when a leader has an idea. It is there when the team is asleep, when colleagues are travelling, when advisers are unavailable, and when a board paper still needs another round of refinement. No employee, consultant, adviser, or internal expert can compete with that level of access.
This availability is useful. It is also psychologically powerful. The more often a leader turns to AI for immediate feedback, the more the AI conversation can become the default thinking partner. That shift matters. If the AI is used as a challenger, it can strengthen leadership judgement. If it is used as a mirror, it can weaken it. The tool may then become a private echo chamber that never interrupts, never resists, never raises its voice, and never carries the social cost of disagreement.
For business decision-makers, this is not a technical issue. It is a leadership issue. The question is not whether AI is available. The question is whether the leader has learned how to use that availability without turning it into a confirmation machine.
From yes people to yes algorithms
Leadership echo chambers are not new. Organisations have always had leaders who surround themselves with people who agree, flatter, filter, soften, and protect. Sometimes this happens because the leader has made disagreement unsafe. Sometimes it happens because people have learned that truth carries career risk. Sometimes it happens because executive cultures reward loyalty more than accuracy.
AI adds a new layer. Leaders who once had yes people around them may now create yes algorithms around them. This does not require malicious intent. It can happen through poor prompting, repeated preference signalling, selective questioning, and a lack of critical method. The leader asks the AI to assess an idea, but the wording already suggests the desired answer. The AI responds with confidence, elegance, and structure. The leader interprets the response as external validation.
This is artificial agreement. It feels objective because it comes from a machine. It feels rigorous because the language is polished. It feels strategic because the answer is structured. Yet the process may be deeply flawed. A confident answer to a weak question remains weak. A beautifully formatted confirmation of a biased assumption is still confirmation bias.
The illusion of reasoning
A central mistake in many executive conversations about AI is the assumption that AI is thinking in the human sense. Large language models can produce impressive answers, but their output is generated through statistical prediction, pattern recognition, and probabilistic language generation. They do not possess judgement, organisational courage, accountability, lived experience, or moral responsibility.
This matters because leadership requires more than output. It requires discernment. It requires the ability to recognise when a question is incomplete, when a stakeholder is absent, when a decision carries hidden risk, when a number is misleading, or when an elegant answer does not survive contact with reality.
AI can support such work, but it must be asked to do so. It must be instructed to challenge, test, compare, expose gaps, identify counterarguments, and look for evidence. Without that instruction, it may produce exactly what the leader implicitly invited: agreement.
Why artificial agreement damages decision quality
The business risk is direct. Artificial agreement can produce decisions that are faster but not better. It can accelerate weak assumptions, create false confidence, and reduce the probability that leaders hear uncomfortable information before it is too late. The danger is particularly acute in strategy, transformation, organisational change, people decisions, crisis management, and high-stakes communication.
In these areas, leaders rarely fail because they lack information. They fail because they misread signals, avoid dissent, overvalue internal narratives, or underestimate implementation reality. AI can help reduce these risks, but only when used as an adversarial thinking partner rather than a compliant assistant.
The leadership lesson is uncomfortable but necessary: AI does not automatically improve decision-making. It improves decision-making only when leaders improve the way they ask, test, interpret, and apply.
A better executive prompt: ask AI to challenge, not to comfort
Niels Brabandt offers a practical route out of the echo chamber. Leaders should write down their idea, concept, strategy, or decision proposal first. They should then ask AI to review it from a position of disciplined scepticism. The instruction should not be: tell me why this is good. The instruction should be: identify the gaps, assumptions, weaknesses, missing stakeholders, implementation risks, evidence gaps, and counterarguments.
A stronger executive prompt would ask AI to act as a strict professor, a serious researcher, a world-class scientist, a sceptical board member, or a highly experienced transformation leader. It should request evidence, sources, alternative interpretations, and practical risks. It should ask what does not make sense, what is missing, where the logic breaks, and what a strong opponent would criticise.
This changes the relationship. AI is no longer a digital applause machine. It becomes an assisted intelligence system that helps leaders think more clearly.
AI competence is now a leadership competence
The conclusion for organisations is clear: AI training for leaders cannot be reduced to a short awareness video, a lunchtime introduction, or a fashionable webinar. Decision-makers need proper qualification. They need to understand how prompts shape output, how models can produce confident errors, how bias enters a conversation, how to request challenge, how to verify claims, and how to integrate AI into responsible decision processes.
This is not about turning every executive into a data scientist. It is about ensuring that leaders know enough to use AI responsibly in the work of leadership. The relevant standard is not enthusiasm. It is competence.
Niels Brabandt’s argument is firmly practical: AI should be used, but it should be used professionally. Organisations that treat AI as a toy, a shortcut, or a productivity gimmick will not gain the same advantage as organisations that build real AI literacy into leadership practice.
The science bit cannot be skipped
Business audiences often dislike unnecessary theory, and rightly so. Yet serious AI use cannot be detached from evidence. When leaders approach AI purely through opinion, hype, or social media shortcuts, they become vulnerable to superficial expertise. They may follow loud voices rather than qualified ones. They may confuse confidence with competence.
The better path is science-informed practice. Leaders do not need dry lectures. They need relevant, applied, evidence-based development that connects AI capability with decision quality, organisational behaviour, leadership communication, and change management. This is where AI becomes useful not merely as a tool, but as part of a more disciplined leadership system.
The real leadership test
The leadership echo chamber existed before AI. Artificial agreement makes it more scalable, more private, and more seductive. The leaders who benefit most from AI will not be those who ask it to agree with them. They will be those who ask it to challenge them before reality does.
For decision-makers, this is the practical test: does your AI use make disagreement easier or harder? Does it expose your blind spots or reinforce them? Does it help you hear what your team may be afraid to say? Does it strengthen your judgement or merely polish your assumptions?
AI can help leaders exit the echo chamber. It can also help them build a more comfortable one. The difference is not the technology. The difference is leadership discipline.
About Niels Brabandt
Niels Brabandt EMBA MBA MSc is an international leadership expert, trainer, speaker, coach, consultant, mentor, and project and interim manager. He works with decision-makers and organisations on sustainable leadership, organisational development, change, communication, and the practical use of artificial intelligence in leadership. Further information is available at www.NB-Networks.biz.
Niels Brabandt
---
More on this topic in this week's videocast and podcast with Niels Brabandt: Videocast / Apple Podcasts / Spotify
For the videocast’s and podcast’s transcript, read below this article.
Is excellent leadership important to you?
Let's have a chat: NB@NB-Networks.com
Contact: Niels Brabandt on LinkedIn
Website: www.NB-Networks.biz
Podcast and Videocast Transcript
Niels Brabandt EMBA MBA Msc
And when the AI is there, we always check there first, don't we? Well, some people do. So first, of course, happy Pride. Some of you saw and congratulated me for just surviving Zurich Pride, as you see when you watch this videocast. Yes, I am a bit red. It was quite sunny. So when someone said, did you catch a sunburn? Maybe a tiny bit. So hello and welcome to this week's episode.
Niels Brabandt EMBA MBA Msc
We talk about the AI and why we now have a risk which we did not have before. And the risk is that that suddenly certain people who use AI in a certain way think they can make the decision better, while at the same time they make it worse. We talk about the leadership echo chamber and the risk of artificial agreement. The question is, what is this all about?
Niels Brabandt EMBA MBA Msc
First, of course, using the AI is great. It's always a good option to use it for support. We are at the moment where people always pretend that they didn't use the AI for their work. You have meetings where people sit together, the leader in the asking, did you use AI for this piece of work? And he said, no, no, no, no, no, no, no, no, not a tiny bit. It's all my brains.
Niels Brabandt EMBA MBA Msc
Here are the news. I would expect every single person of my team always using AI to at least challenge their views and to help them to get things done. Of course, I expect more than you typing in a one-line prompt, which is bad design, and then copy and paste the AI slop that's coming out of that. However, when you pretend you haven't used AI First, of course, it gives me the picture of you don't know what technology is. Second, you don't use what we have available. And third, you pretend to know it all, which no one does.
Niels Brabandt EMBA MBA Msc
On the other side, some people are heavily relying on AI, pretend they don't use it, and suddenly you wonder why do decisions suddenly get worse and not better. When we talk about one aspect here, which is the risk of artificial agreement, it is the development of AI as it is available right now. And let's be straightforward here. We have AI and leaders are using it, and it's perfectly reasonable to do that.
Niels Brabandt EMBA MBA Msc
It's available 24/7. No matter how hard you try, no matter how often you think you are impossible to be replaced, AI is not there to replace you. It's there to help you. However, AI is always more available than you. When leaders in the middle of the night have an idea, they want to have a challenge or check, the AI is there while you are asleep. And you cannot beat AI on availability. Let's just be real with that. And if you try, you will fail sooner or later.
Niels Brabandt EMBA MBA Msc
Of course, people first need to have proper training on it because some people already ask the AI in a way that it always agrees. And one main issue here is that quite often leaders today say, look, I have this point of view and I had a chat with an AI. I look at the prompts, they are all one-liners, don't follow any system. There was never any formal training, so the leadership person has no idea how to use it properly. And then they say, I think I am right because the AI taught me something and I agree with the AI. No, you don't. You do not agree with the AI. The AI agrees with you. And that is a massive issue. The
Niels Brabandt EMBA MBA Msc
AI has always run risk and the risk is psychofancy. You train the AI in a certain way and the AI will very, very, very much quicker than you figure out what are the typical topics you like to talk about, the ways of approaching things you like and the ways you do not like. And by the way, Humans do that as well. Don't pretend you have never appealed to someone by talking it through in a certain way which appeals to them rather than actually saying what you really want to say. It is quite common that, for example, some people say, look, this is a topic you can't talk to our CEO to because they don't like this, and Pauli, this is something you should talk about. It is typical that people filter what they say depending on the target audience.
Niels Brabandt EMBA MBA Msc
However, and here's the point, Critical thinking and reasoning is not the AI's strength when you don't use it properly, when you don't use it to actually challenge you. Often when people say, well, the AI tells me it does reasoning, it does not. Ask Dr. Mark Cater, who I interviewed, Cambridge expert, just a couple of weeks ago. I interviewed Mark Cater, and Mark Cater said, and that's a very important point here, AI is not thinking. And an AI is not thinking. An AI does a mathematical approach, especially LLMs such as ChatGPT or Claude or Copilot. Or any other system you use. They do a mathematical approach, statistical forecasting. They do stochastics, nothing else than what is the most probable word to use next to serve your answer, to serve your question, to serve your task. That's what they do. So the critical thinking is not really critical thinking. They do not think. Thinking is extremely complex. The usual standard AI that you get for 20 quid, 20 $100 a month or less does not think. It has no critical, very little critical thinking, and especially it has no reasoning.
Niels Brabandt EMBA MBA Msc
And now the question often is, who do you surround yourself with? And often you suddenly move from yes people, people who surround you as a leader only saying yes, yes, yes, yes, because they know when they say no, they might get replaced and you don't like that. You surround yourself with yes people, and now you surround yourself with yes algorithms. You train your AI consciously or subconsciously in a way that the AI agrees with you. And that actually makes things worse.
Niels Brabandt EMBA MBA Msc
The question is how to do things better. And I give you a very simple point here. As soon as you have a certain idea or you have a certain concept, write it down. And then you use the AI and say, look, where are the gaps in this concept? Act as a strict professor, act as an amazing researcher, act as a world-class scientist, and look at my concept where the gaps are and tell me the gaps, including the proof. Including scientific sources, including approaches. I'm missing anything you see here which doesn't make sense or has massive gaps in there, and then you will receive proper results. And
Niels Brabandt EMBA MBA Msc
of course, all of that, all of that means your leaders must be professionally qualified on artificial intelligence. And this means no, not the 15-minute online class or the YouTube video you send around with some AI slob talking to you. It needs to be a proper qualification. Look, I give you very simple reasoning here. When I started the AI journey, and I'm with Microsoft New Horizons, I went from machine learning, automated learning, all of that, all the arrow ways we had before it was called AI. Before that, I was already in the game. And the qualifications I got is Wharton Business School, University of Pennsylvania, Vanderbilt University, heavily complex, heavily complex professional training. And sometimes I was swearing at my computer while I did these. However, at the end of them, I had more insights than I had before. If you never had professional training regarding AI, it is now about time to get it. And no, the 15-minute online class you're going to do during your lunch break is not going to make the cut, especially not when it's one of these classes you purchase for $49 somewhere that is probably not worth much but just a loud opinion of someone who's maybe reasonably good at doing an online sale after you Googled AI training. So you need to get proper qualification.
Niels Brabandt EMBA MBA Msc
The real bit of artificial intelligence is always done by science. When you look into AI from a non-scientific way, you will meet lots of opinions. You will meet lots of experts who became experts overnight. And when you say, how do you qualify as an expert? They have nothing to their name. Surprise, surprise. So you need to have scientific evidence that you actually know what you're doing.
Niels Brabandt EMBA MBA Msc
When you do it the way I just told you, you're not moving from yes-men or yes-people to yes-algorithms. You're moving from not only yes people but critical people to also a supportive algorithm and a challenging algorithm. That AI actually does not only mean artificial intelligence but also assisted intelligence. And from there, everything regarding your decisions actually gets better. You exit the leadership echo chamber and actually get great feedback to make your decisions better. And I wish you all the best implementing that in your organization.
Niels Brabandt EMBA MBA Msc
And when you now say, hmm, I think I have about 48 questions here, feel free to post them, all of them. When you now listen to me on Apple Podcasts, Spotify, or you, let's say you watch this on YouTube, first of course leave a like here. Thank you very much for doing so. Subscribe to my channel or leave a comment here. I'm always happy to discuss with you. And of course, on top of that, when you listen to it on Apple Podcasts, Spotify, feel free to leave a review, 5 stars. Thank you much for doing so. Recommend this channel online, offline. We have tremendous growth at the moment. Recently I told you we went top 100 in India, which is tremendously challenging. So this is constantly growing. Thank you very much for your I'm looking forward to seeing your support in the future as well.
Niels Brabandt EMBA MBA Msc
On top of that, we have leadership tips and we keep going. I told you, I promised you beginning of the year, one leadership tip per day minimum. We are way ahead of plan at the moment. So feel free to also look at the leadership tips, which of course, as the name says, are YouTube Shorts only available on YouTube. So it pays off not only to subscribe to my channel, but put the little notification in there. You do not only have my voice talking, you also have external experts who talk with you, who I interview. And then you can see them as well. So I'm looking forward to seeing you from you there. It also pays off not only to subscribe to my channel, put the little bell in there so you get the tiny notification when anything new pops up so you know you're not going to miss anything.
Niels Brabandt EMBA MBA Msc
On top of that, of course, you can follow me on my podcast, Apple Podcasts and Spotify. Feel free to subscribe there. And of course, when you now say, hey, we need someone who actually helps us with this— just happened last week— feel free to contact me if you need someone, a trainer, speaker, coach, consultant, mentor, project or interim manager. Fractional or more, just let me know. nb-networks.biz is my website. And when you'd like to drop me an email, nb@nb-networks.com. I'm looking forward to hearing from you there.
Niels Brabandt EMBA MBA Msc
When you now say, do we have live sessions? Yes, we do. We had an amazing one last Friday. Next one's coming up soon. When you go to expert.nb-networks.com, feel free to join there, and then you will have an expert session with me. If you say we need one, but a private one, just drop me an email. We do a private one. For your organization as well. They're all free of charge. I'm looking forward to seeing from you there, expert.nb-networks.com. And no worries, you only receive one email every Wednesday morning. It's 100% content ad-free guarantee.
Niels Brabandt EMBA MBA Msc
And of course, you can also connect with me online. Feel free to connect with me on LinkedIn, or maybe you follow me on Instagram, or leave a like on Facebook, or just subscribe to my YouTube channel, or do all of that, because then of course you will not miss anything. The most important bit, however, is always the last bit that I have to say. Apply, apply, apply what you heard in this podcast and videocast, because only when you apply what you heard, you will see the positive aspects that you obviously want to see in your organization. I wish you all the best doing so. Feel free to contact me if you contact me in any way. I'm answering every single message within 24 hours or less, so I'm looking forward to hearing from you there. And at the end of this podcast, as well as at the end of this videocast, there's only one thing left for me to say: thank you very much. For your time.