Sérgio Tavares
Agency is the ability technology should give us to understand what we are doing, and allow us to do what we intend to.
In most companies, the most forward-thinking people are told to “push back.” This is often a designer’s cultural role: to question the status quo, to fight inertia. That instinct now runs into the AI hype cycle.
It’s common to see crowds of AI snake oil being sold on LinkedIn, so a common reflex becomes rejection, in an attempt to “think differently”.
The irony is that designers (who fought hardest when Photoshop, Figma, or Webflow shifted the field, or when Microsoft Teams replaced Slack or Google Meet) are now gatekeeping against the tools that might change it again.
Admittedly, this response is well warranted. Tech companies are among the biggest evildoers of our time, and AI (in its superintelligence form) does pose an existential risk on us. But that is different from using AI to help with your Bolognese recipe, your workshop ideas or automating work processes (even creative ones). We may even dispute copyrights, and still there would be a million ways in which AI can be a good tool for at work.
Why it matters: When resistence becomes a posture instead of analysis, teams waste energy fighting shadows while missing real point.
The danger of saying no before thinking
It’s very common that meetings about AI (and edge projects, for that matter)
The job is not to embrace nor reject AI wholesale. The job is to resist herd belief—belief formed by hype and fear, rather than by tested evidence. Nilsson reminds us that beliefs guide action but must remain tentative and changeable (Nilsson 2014, 4).
Complexity science shows how herd cascades amplify early signals into market manias. Fear of skill erosion exaggerates risk perception, and that becomes a powerful tool of resistance. And here’s a little example I love, the The Simple Sabotage Field Manual by the CIA, with tactical hints on how to sabotage an organization’s productivity. It was declassified in 2008, and frequently resurfaces among corporate discussions.
So when it comes to resistance to change, here’s one gem:
Together, beliefs, herd cascades and either ambition or fear turn these forces turn “strategic thinking” into blind embrace, and “critical thinking” into “blanket rejection.” And neither is good.
The result is a noisy polarization, because in social media (like LinkedIn) there is selective exposure: “people choose and share confirming content” (Modgil 2021). LinkedIn amplifies this, since most of us are there to do business, thus we seek visible consensus (likes, comments), and discourage dissent.
Three things to consider about the hype cycle of AI
So, now that we have considered a simple attitude towards innovation (being tentative and changeable) and considered the context we are in (the disillusionment phase of the AI hype cycle), here’s a few other points that explain more polarized attitudes:
- Herd belief is like contagion. Herd behavior operates like epidemic transmission: once a few signal rejection (or hype), the belief spreads by social proof. Some people pride themselves on independent thinking, but they are as just as susceptible. Nilsson notes that many beliefs come not from direct evidence but from what we read, hear, or infer socially (Nilsson 2014, 21–27). If there’s fear mongering, this escalates even quicker.
- The hype cycle trap. Gartner’s model (above) shows the rise from inflated expectations to disillusionment before stabilization. Those who join the “AI is trash” chorus risk anchoring the field in the trough, blind to productive adoption. In a very similar way to the Monopoly man, who won’t see how AI should be used, and when.
- Fear as skill protection. Early resistance came from a place of insecurity: if a machine can generate a layout, what’s left of my craft? But protecting craft by denying tools confuses medium with mastery. Critical design depends on framing problems, not just producing outputs.
As we rise to a plateau of stable productivity, observe the curve of the wave. The opportunity is to be the surfer.
Design the path
That is not to say that a great dose of critical thinking isn’t good. It it a vital part of the process. In order to preserve agency, we need to live by clear guidelines. And in this context, it means giving teams the capacity to separate the thing (AI) from the herd noise around it. That requires:
- Principles: Resistance is useful when it punctures hype, not when it rejects experimentation.
- Patterns: Trial labs, safe sandboxes, critical showcases of where AI fails and where it helps.
- Why, When, How : Have a structured decision flow: when to use AI, when not, and how to escalate judgment calls.
- Governance: Treat AI decisions like design critiques: evidence, iteration, and reversible paths.
A quick cheat sheet
Who | What they’ll say | What to do |
Developers | “Nope.”
Will plead for clean code, concerns with integration complexity, scalability prices and risks in security | Listen carefully. Devs aren’t by nature co-creative. They often assess with computational precision, as fast as they can (it’s a symbol of status). Listen, and flip the narrative: what would have to be true in order for this to work? |
Legal | “Maybe not.”
They will push for risk-aversion and a strong legal backbone. | Seek the resources to comply. You will need tightly scoped pilots and complete disclaimers stating clearly what and what not to expect from the AI output. |
Brand | “Yes, but…”
They will fear bad optics, controversial uses or irresponsible use. | Draw the lines.
Frame uses as internal use first, external later, and human-in-the-loop until further notice. |
Designers | “Is this even good?”
They will inquire if this is ethical, sustainable and safe. | Co-create.
Engage in the discussion, welcome questioning, and seek answers. Then, co-create better alternatives. |
Mgmt | “Faster!”
They may want early demos to lock new business, and accountability plays a role here. | Manage expectations with real KPIs.
Work with a simple MVPs and clear KPIs. Deliver on a clear promise, making sure the big future of the project is visible. |
References
- Nilsson, Nils J. 2014. Understanding Beliefs. MIT Press.
- Banerjee, Abhijit. 1992. “A Simple Model of Herd Behavior.” Quarterly Journal of Economics.
- Gartner. 2023. “Hype Cycle for Emerging Technologies.”
- Arthur, W. Brian. 2013. Complexity and the Economy.
- Furedi, Frank. 2006. Culture of Fear Revisited.
- Modgil, S. et al. 2021. A Confirmation Bias View on Social Media Induced Polarisation During Covid-19.
- Office of Strategic Services. 1944. Simple Sabotage Field Manual. Washington, DC. Declassified by the Central Intelligence Agency, 2008.