Why It’s Getting Harder to Hold Nuanced Positions
Influence in the Age of AI
Many people sense that public disagreement has changed. Positions harden quickly. Conversations polarize even when stakes are low. Letting go of a view feels costly in a way it did not before.
This shift is often attributed to politics, media incentives, or declining civility. Those explanations capture part of the picture, but they miss something more basic: beliefs are increasingly doing the work that stable identity once did.
When that happens, nuance becomes difficult to sustain.
When Positions Start Carrying Identity
In the past, disagreement was often buffered by external structures. Work, community, and institutions provided enough stability that people could revise opinions without threatening their standing, relationships, or sense of self.
Many of those buffers no longer function the same way.
Workplaces once offered clearer roles, longer time horizons, and shared standards for competence. Communities provided continuity, shared rituals, and face-to-face correction. Institutions supplied reference points for legitimacy and authority, even when people disagreed. Families, churches, unions, and professional bodies helped separate belief from belonging.
Those structures have weakened, fragmented, or been replaced by systems that reward visibility, speed, and alignment over judgment.
As a result, disagreement now carries more personal risk.
When fewer external structures hold identity steady, people rely more heavily on internal markers—beliefs, positions, affiliations—to maintain coherence. A change in view no longer feels contained. It spills into questions of credibility, belonging, and self-trust.
This is why positions harden.
It is not because people suddenly prefer certainty. It is because changing one’s mind now costs more than it used to, socially and psychologically.
Why This Produces Sides Instead of Judgment
Nuance requires slack. It depends on time, shared context, and confidence that disagreement will not result in exclusion or loss of legitimacy.
When those conditions erode, several shifts follow:
Binary frames feel safer than conditional ones.
Alignment feels more stabilizing than inquiry.
Defense replaces evaluation because retreat carries social and psychological cost.
People are not becoming less capable of thought. They are operating in environments where maintaining coherence—keeping a stable sense of self, credibility, and belonging—takes priority over revising conclusions.
This helps explain why debates so often collapse into sides, even when the subject matter does not warrant it.
Influence Without Persuasion
Another change accompanies this shift. Influence increasingly operates without needing to persuade in the traditional sense.
In an earlier piece, I described the human–AI salience loop: systems designed for engagement repeatedly surface certain signals, reward attention to them, and feed that attention back into future selections. Over time, what is shown more often begins to feel more relevant. What feels relevant draws more attention. The loop reinforces itself.
That loop matters because engagement does not require agreement. It only requires exposure.
In the age of AI, this process becomes more efficient.
Large-scale behavioral data allows systems to model patterns of attention, response, and susceptibility across populations and individuals. Probabilistic models do not need to know what a person believes. They only need to estimate what is likely to hold attention, when reinforcement is most effective, and which signals are most likely to be returned to the user next.
User feedback—clicks, pauses, scroll depth, repetition, abandonment—feeds back into these models continuously. The system adjusts in real time, maintaining engagement by refining what is made salient rather than by persuading directly. The result is a system that shapes attention patterns first, allowing confirmation bias and familiarity to carry beliefs forward without requiring explicit agreement.
Instead of arriving as a single argument or message, influence now accumulates through repeated encounters across everyday systems—work tools, information feeds, entertainment, social platforms, health decisions, and routine digital interactions. No single signal feels decisive. Most feel neutral or even helpful.
What changes first is not belief, but salience—what draws attention, what comes to mind easily, what feels normal enough to defend. Once certain ideas or frames are repeatedly made salient, confirmation bias begins to do the rest.
Repeated exposure matters even when people disagree with what they are seeing.
Humans are social animals. We learn what is safe, relevant, and worth responding to by observing patterns in our environment and in each other. Familiarity lowers cognitive effort. Lower effort increases uptake.
Over time, repeated exposure shapes what feels reasonable to act on and what behaviors feel normal or defensible.
This is where engagement-driven systems intersect with identity. Reinforced signals support coherence by helping views feel internally consistent and socially shared.
The process is gradual. It rarely feels coercive.
By the time beliefs harden, persuasion is no longer required. The work has already been done through attention, repetition, and the unconscious consolidation of what feels thinkable.
A Condition Worth Examining
When belief starts carrying identity load, disagreement feels existential and nuance feels risky. That does not mean people have lost the capacity for judgment. It means the environment has raised the cost of using it.
The outcome does not require an intention to destabilize. When influence systems reward disruption, novelty, and emotional response more reliably than stability or resolution, erosion of shared reference points becomes a predictable result.
That terrain is not abstract. We are already standing in it.
In the next post, I’ll introduce a model that makes this accumulation visible—not to persuade, but to give readers a clearer way to judge what they are already experiencing.
I am a UX designer and researcher working at the intersection of AI, human cognition, and systems design. My work develops design strategies that preserve judgment, learning, and cultural resilience as AI becomes embedded in everyday life. You’re welcome to connect with me on LinkedIn and subscribe here for updates on my two-book series on sustaining human intelligence and building a resilient culture in an AI-driven world.