A slide that smells like AI is your deck’s emoji-filled LinkedIn post. In 2026, it’s the fastest way to lose attention in a quarter of a second.
Key Takeaways:
• Workslop = AI-generated work that looks acceptable but is empty. In a presentation, the effect is multiplied because it breaks the trust in real time between the speaker and the listener.
• The brain is becoming an anti-slop radar. We recognize AI at the perceptual level, below the threshold. It’s a reflex, not a decision, and it’s getting better month after month.
• A slide has a quarter of a second to not be labeled as AI slop. If the audience even suspects, the attention span collapses and does not recover.
• A presentation exists to build trust, not to convey information. If it breaks trust, it has failed, even if the contents are right.
• AI should be used upstream, never downstream. Thinking with AI: yes. Generating the final slide with AI: zero respect for those in front of you.
Think about the last time you scrolled through LinkedIn. At one point you passed over a post full of emojis, perfectly symmetrical sections, sentences with the rhythm of a ChatGPT left free to write without a soul inside.
How long did it take you to scroll any further? Half a second? A second?
You haven’t read. You didn’t rate. You didn’t think. Your brain recognized the pattern, pulled out the “AI slop” label, and authorized you not to invest attention. Game over.
Now transfer exactly this scene into a meeting room. You are in front of fifteen people. Open the first slide. And that slide smells the same as that post.
What happened to you on LinkedIn, is happening to them in front of the projector. Same quarter of a second, same label, same mental shake towards elsewhere.
The brain is becoming an anti-slop radar. And it is improving quickly.
Two years ago, a text written by AI seemed magical to us. Today we recognize it at a glance. The same goes for images, videos, emails with an overly orderly structure, bios with three symmetrical emojis and the motivational aphorism.
We are not learning only rationally. We are learning on a perceptual level, below the threshold, before we think. It is a reflection, not a decision.
And like any reflex, it’s getting faster and more reliable month after month. In twelve months it will be even sharper. In twenty-four, it will be a seismograph. This does not stop.
People don’t need to conclude that your slide is made by AI. They just suspect it for a moment, and you’ve lost your room.
In a presentation, the workslop effect is twice as dangerous
In the other deliverables, the workslop costs time. A vague email, a smoky report, a slide deck sent by email and never opened: the damage is process. It is recovered.
In a presentation, no. Because in a presentation there is something that is not present in other deliverables: the real-time relationship between the speaker and the listener. A relationship that is based on a single material: trust.
A presentation, after all, is an explicit request. You’re asking people to give you a gift: their time and attention. They are not renewable resources. They are over, scarce, disputed. And you’re asking them now, in this room.
The implicit pact is simple: “Pay attention to me because I have paid attention to you even before I enter here”.
If the first quarter of a second of your first slide communicates that you didn’t do it, that the slide was generated without you looking at it, without you choosing it, without you thinking about it, that pact breaks. And it breaks in a way that is difficult to recover, because it breaks at the perceptual level, before the rational part of the brain has time to intervene.
If you have to take away a single sentence from this article, it is this one.
The quarter of a second that decides everything
In Neuro Presentation Design I have written entire pages on this idea, because the presenter always underestimates the window of time in which the audience decides whether to listen to you or not.
It’s not the first five minutes. It’s not the first thirty seconds. It is the first quarter of a second since a slide appears.
It is in that frame that the audience’s brain classifies the signal: is it worth paying attention, or do I relax, pick up my phone, look at my watch, think about myself?
That quarter of a second is not managed by the rational part of the brain. It is managed by the perceptual system. You can’t bargain with it. You can’t convince him. You can only give him a signal that he deserves to stay awake.
A slide that looks AI slop sends the opposite signal. It sends the signal that those who made it did not invest at that time. And if he didn’t invest, why should I?
A small confession, from someone who lives on stage
In recent years, between events, boards, lectures, executive training, I have brought thousands of slides to projection.
To date, I have never put a single AI-generated slide in front of an audience.
Not out of fundamentalism. Every time I thought about doing it, I realized that it was a lack of respect for those in front of me.
It would have meant asking my audience to invest their attention in something that I, for one, had not invested mine in.
It would have meant presenting myself to that pact as a defaulter. And a pact signed by one party does not work.
The moment you stop investing in your slides is the moment you can no longer ask anyone to invest in you.
So how do you use AI in presentations?
I’m not saying don’t use AI when building a presentation. That would be absurd, and it would be against everything I’ve been preaching for years about working lean.
I’m saying something else. AI must be upstream, never downstream.
Upstream: thinking with AI. Exploring angles, stress-testing the thesis, doing research, structuring the flow, rewriting a text that doesn’t run, generating alternatives, dismantling objections, creating the zero version of a storyline.
Downstream, that is, on the slide that ends up in front of people’s eyes, there must be your decision. Every word chosen. Every piece of data verified. Every chart rethought. Each image weighed.
The rule of thumb is this: if you can’t defend every millimeter of your slide because you haven’t actually looked at it, that slide isn’t ready to be projected. It’s ready to stay on your drive, or worse, in the sandbox of the LLM that generated it.
Having AI generate slides is a return to Fordism: push production, waste upstream, quality control downstream that tries to save what can be saved. Just-in-time overturned that model starting from the customer’s request and pulling production backwards, without stocks. A lean presentation does exactly that: it starts with the audience in front of you and pulls back each slide.
The upstream AI amplifies that thought. Downstream AI replaces it with a push of average outputs.
Why curating slides in the age of AI sets you apart right away
There’s also a positive side effect: the more the world fills up with hastily generated, superficial, or poorly curated content, the more apparent it becomes who takes the time to actually think.
You can immediately notice when a slide has been carefully built. When an argument has been chosen, not just assembled. When a citation has been verified. When a piece of data has been understood, discussed and transformed into a clear message.
In general noise, these signals shine much brighter than before. Those who present with care today are not only presenting well. He is declaring, even with just the first slide, that he is among those who still put his head into it.
It is an act of positioning, even before an act of communication.
The real function of a presentation is not what you think
A presentation does not exist to convey information. For that there are PDFs, emails, reports, databases. A presentation exists to build a relationship of trust between you and your listeners, in real time, in that room, in that moment.
If it doesn’t build trust, it has failed. Even if the contents are right. Even if the numbers are correct. Even if the storyline holds.
And trust, today, begins with a quarter of a second in which the audience decides whether they have you in front of them as a professional who has invested in them or as a professional who is wasting their time with something they generated five minutes earlier.
An AI-generated slideshow is your deck’s emoji-filled post. Don’t post it. Don’t project it. Don’t sign it.
Conclusion
AI isn’t killing presentations. He is unmasking them. For the first time we have a tool that, if used badly, makes visible at a glance those who have not invested in their work. And he does it in a quarter of a second, in front of the whole room.
The choice is not between using or not using AI. It is between using it upstream, to think better, or downstream, to fill slides for you. In the first case, it is the greatest thought amplifier we have ever had. In the second it’s the fastest way to tell your audience that you didn’t care enough about their time.
When you walk into a room with fifteen people in front of you, each of them is giving you a gift. Your first slide is the first thing that says if you are treating that gift as such, or as something owed. Everything else in the presentation is played on that first signal.
In 2026, taking care of slides is not aesthetic. It is respect. And it’s the most underrated competitive advantage you have.
FAQ – Frequently Asked Questions
What is workslop?
Workslop is the term brought into focus by Harvard Business Review to describe AI-generated work that seems acceptable on the surface but is devoid of substance. Documents, reports, Excel files, presentations, and emails that look like a job well done, but lack the thinking needed to move a project forward. It shifts the cognitive load from the producer to the receiver.
Why does the workslop have a greater impact in presentations?
Because in presentations there is a variable absent in other deliverables: the real-time relationship between speaker and audience, which is based on trust. A slide that seems AI-generated breaks that pact in a quarter of a second, and at that point the attention is no longer recovered, even if the content is valid.
How does the brain recognize an AI-generated slide?
Not on a rational level but on a perceptual level. The visual system identifies patterns of symmetry, textual rhythm, stylistic choices that it associates with the output of a generative model. This reflex is rapidly refining: each month of exposure to AI content makes the audience more capable of intercepting the slop below the threshold.
So you can’t really use AI to make slides?
AI can and should be used in the preparation phase: brainstorm, thesis stress-test, research, generation of alternatives, rewriting texts, building a zero version of the storyline. What doesn’t work is having the AI generate the final slide that will be under the eyes of the audience. AI upstream, never downstream.
What does AI fluency mean?
AI fluency is the ability to work with AI in a structured way: knowing when AI expands thought, and knowing when it is replacing thought. More than the ability to use AI, it’s the ability to know when not to use it. It is the heart of the framework described in the book AI Fluency.
What do you mean by “the quarter of a second”?
It is the time that the perceptual system of the audience’s brain takes to classify the first contents of a slide as worthy of attention or not. In that frame there is no reason: there is only pattern recognition. A slide that looks like workslop immediately sends the wrong signal. The concept is explored in depth in the book Neuro Presentation Design.
How do you build trust with a slide?
With signs that demonstrate investment: a thoughtful visual composition, a chosen and not automatic title, verified data, a clear hierarchy, consistent images, the absence of redundancy, an explicit choice of what to leave out. All elements that AI alone does not produce: it produces average and symmetrical versions of everything.
Comments on AI is killing your presentations. And your audience notices it before you do.