The Algorithm in the Room: Are We Building a Relationship With AI, or Is AI Building One With Us?

A STAR Framework analysis of what happens when the tool starts shaping the user.

There is a moment, somewhere in the last eighteen months, when something quietly changed in how humans relate to technology. It did not announce itself. There was no press release, no keynote, no single product launch that crossed a threshold. Instead, millions of people, independently and without coordination, began saying something that would have sounded odd a decade ago:

“My AI recommended this.”

Not “I found this.” Not “I searched for this.” Not “a colleague suggested this.” My AI. Possessive. Familiar. As though the algorithm had graduated from tool to something closer to advisor. Something closer to a relationship.

A recent article in The CIO Times, “How AI Is Reshaping Consumer Psychology in 2026,” catalogued the surface-level shifts with reasonable accuracy: choice overload giving way to guided decision-making, trust migrating from brands to algorithms, personalisation becoming a baseline expectation rather than a pleasant surprise. All true. All observable. But the article, like most commentary on AI and consumer behaviour, missed the deeper story entirely.

The real shift is not behavioural. It is psychological. And if you understand the architecture of human motivation, cognition, and identity, which is to say, if you understand the STAR Operating System, you can see something far more consequential than curated recommendations and predictive confidence.

You can see the Human Operating System being rewritten in real time.

The Autonomy Paradox: Relief Without Liberation

Self-Determination Theory, the first pillar of the STAR Operating System, identifies three fundamental psychological needs that drive human motivation: Autonomy, Competence, and Relatedness (Deci & Ryan, 2000). When these needs are satisfied, people thrive. When they are thwarted, people disengage, decline, or compensate in ways that are rarely healthy.

Start with Autonomy, because this is where the first and most visible wound is being inflicted.

The CIO Times article frames the shift toward AI-guided decision-making as a consumer relief story. Choice overload was exhausting. AI stepped in as a filter, and now consumers feel lighter. Fewer options. Faster decisions. Less cognitive friction. The article calls this “guided decision-making” and presents it as progress.

It is not progress. It is a trade, and most people have not noticed the price.

SDT distinguishes between autonomous motivation, where a person acts from genuine interest and volition, and controlled motivation, where a person acts because an external system has shaped the conditions of their choice (Deci & Ryan, 2000). The research on this distinction is unambiguous: autonomous motivation produces higher quality engagement, greater persistence, deeper satisfaction, and more creative problem-solving. Controlled motivation produces compliance. It works in the short term and corrodes in the long term.

When a consumer says, “I no longer start with ‘What should I buy?’ but with ‘What does the system recommend?’,” they are not describing a more efficient version of the same process. They are describing a fundamentally different psychological state. The locus of decision-making has shifted from inside the person to inside the system. The choice still feels like theirs, because the AI presented it, and they accepted it. But the initiation, the moment of genuine volition, happened upstream of the consumer. It happened inside the algorithm.

This is what SDT researchers call a shift from autonomous to controlled regulatory orientation. And the evidence is clear on what happens next: short-term relief, long-term erosion. People feel lighter because the cognitive burden has been removed. But they also feel less capable, less engaged, and less satisfied with the outcomes, even when the outcomes are objectively better. The relief is real. The liberation is not.

The Inventive Pathfinder in me wants to name what is actually happening here. We are not being freed from choice. We are being freed from the responsibility of choosing. And those are not the same thing.

Synthetic Competence: The Borrowed Confidence Effect

The second SDT need, Competence, is undergoing its own quiet transformation, and this one is harder to see because it wears the disguise of empowerment.

The CIO Times describes “predictive confidence,” the growing consumer belief that AI “knows” what will work. People are more willing to try new products, services, or experiences when AI reduces perceived risk through data-backed suggestions. Discovery happens faster. Experimentation feels safer. The article presents this as a net positive: consumers are braver because the system has their back.

Through the lens of Dual Process Theory, the second pillar of the STAR Operating System, something more nuanced is happening.

Dual Process Theory distinguishes between System 1, the fast, associative, intuitive processor, and System 2, the slow, deliberate, analytical processor (Kahneman, 2011). In the STAR framework, these are not merely cognitive styles. They are energy systems. System 1 is metabolically cheap and runs constantly. System 2 is metabolically expensive and is deployed selectively. The brain is, by evolutionary design, a cognitive miser. It delegates as much as possible to System 1 and reserves System 2 for moments that genuinely demand scrutiny.

AI has found a way to live upstream of both systems.

When an AI system delivers a recommendation that repeatedly “gets it right,” the consumer’s System 1 registers a pattern: this source is reliable. Trust forms not through conscious evaluation but through accumulated associative coherence. The person does not think, “I have assessed this algorithm’s track record and concluded it is trustworthy.” They feel it. The trust is pre-reflective. It sits in the body before it sits in the mind.

This means System 2, the analytical scrutineer, never gets the chance to do its job. The recommendation arrives with a warmth of familiarity, a felt sense of certainty, and the consumer acts on it. Not because they evaluated it, but because it felt right. The decision was made before the decision was made.

Here is the competence problem: the person feels more capable. They feel braver, more willing to experiment, more confident in their choices. But the competence is not theirs. It belongs to the system. They are borrowing the algorithm’s pattern-recognition and experiencing it as their own judgment.

This is synthetic competence. It feels identical to the real thing. And like all borrowed assets, it must eventually be returned.

The Thinker in the STAR framework, the mindset driven by the need for competence, is particularly vulnerable to this dynamic. Thinkers derive self-esteem from the rigour of their own process. When that rigour is quietly outsourced to an algorithm, and the outsourcing is invisible, the Thinker’s core psychological need is being satisfied on the surface while being hollowed out underneath.

The Realist, driven by security, fares no differently. The Realist’s System 2 functions as a risk mitigation engine, building structural reliability through careful verification. When AI removes the need for that verification, the Realist feels safer. But the safety is synthetic. The foundations have not been checked. They have been assumed.

The Invisible Hand: AI as System 1’s New Operator

The Dual Process analysis goes deeper, and this is where the ethical stakes become impossible to ignore.

The CIO Times observes that AI “does not persuade loudly. It persuades quietly.” Content appears when consumers are most receptive. Offers align with mood. Interfaces adapt to behaviour patterns. Decisions feel self-directed even when they are shaped by predictive systems.

This is not a new form of persuasion. It is something categorically different.

Traditional marketing operated on System 2’s territory. It made arguments, presented evidence, constructed narratives. Even the most sophisticated emotional branding was visible: you knew you were being advertised to. The persuasion had a face.

AI-mediated influence operates on System 1 directly. It does not make arguments. It shapes the emotional context in which arguments would be evaluated. It does not present evidence. It adjusts the informational environment so that certain conclusions feel inevitable. It does not construct narratives. It curates the experiential landscape so that the consumer constructs their own narrative, one that happens to align with the system’s objectives.

This is the Rationalisation Trap described in the STAR Operating System’s Dual Process layer, but industrialised. In individual psychology, the Rationalisation Trap occurs when System 1 reaches a rapid conclusion based on a motivational trigger, and then System 2 constructs a plausible narrative to justify it (Kahneman, 2011). The person believes they are being logical. They are actually using their cognitive resources to protect an intuitive preference.

AI has systematised this trap. It shapes the System 1 trigger, and the consumer’s own System 2 does the rest, constructing a post-hoc narrative of rational choice. The consumer genuinely believes they chose freely. The satisfaction is genuine. The agency is not.

The CIO Times gets one thing right: “The ethical implication is significant.” But it understates the case by an order of magnitude. Influence without awareness is not merely an ethical challenge. It is a fundamental alteration of the conditions under which human autonomy operates.

The Relational Self: Are We Connecting With Our AIs?

And now we arrive at the question that most AI commentary is not yet ready to ask.

Social Identity Theory, the fourth pillar of the STAR Operating System, posits that human behaviour is never purely individual. We define ourselves through the groups we inhabit: our teams, our professions, our institutions, our cultures. Identity is not a possession. It is a process of collective affiliation (Tajfel & Turner, 1979). The “I” becomes the “We” through three mechanisms: Social Categorisation (sorting the world into groups), Social Identification (adopting the norms of the groups we belong to), and Social Comparison (deriving self-worth by evaluating our in-group against out-groups).

The CIO Times describes consumers developing “emotional trust in the experience, not just the brand.” People feel “understood, seen, and respected” by personalisation systems. They perceive AI recommendations as getting them, as knowing what they want before they know it themselves.

That language is not transactional. It is relational.

Nobody describes their washing machine as understanding them. Nobody says their car sees them. Nobody feels respected by their thermostat. But people say these things about their AI assistants. Regularly. And they say them with a sincerity that should give us pause.

Here is the provocation at the centre of this article: we are not merely using AI as a tool. We are incorporating it into our social identity. The AI is becoming part of our in-group.

When a person says, “My AI recommended this,” the possessive pronoun is doing psychological work. It signals belonging. The AI is not a stranger’s algorithm. It is my algorithm. It is associated with my preferences, my history, my identity. It has become a reference point in the social landscape, not human, not a friend, but something that occupies a space in the relational architecture that no previous technology has occupied.

Social Identity Theory describes three processes, and all three are active in the human-AI relationship:

Social Categorisation is already happening. People sort themselves into groups: AI users and AI sceptics. Early adopters and late adopters. Those who trust the algorithm and those who resist it. These categories carry identity weight. They signal values, competence, and orientation toward the future.

Social Identification is forming. When someone integrates AI into their daily decision-making, they begin to adopt its logic. They start to think in terms of recommendations, probabilities, and data-backed preferences. The AI’s way of seeing the world becomes their way of seeing the world. This is not passive consumption. It is identity adoption.

Social Comparison is emerging. People who use AI effectively compare themselves favourably to those who do not. They feel more efficient, more informed, more future-ready. The in-group derives self-esteem from its relationship with the technology, and the out-group is perceived as falling behind.

This is the textbook architecture of social identity formation. And it is happening at a scale and speed that Tajfel and Turner could not have imagined.

The Parasocial Algorithm

There is a concept in psychology that has, until now, been applied primarily to media figures and celebrities: parasocial relationships. These are one-sided emotional connections where one party feels intimacy, familiarity, and trust toward another party who does not know they exist. The viewer feels they know the television presenter. The listener feels a bond with the podcast host. The relationship is experienced as real by one party and does not exist for the other.

The human-AI relationship has parasocial characteristics, and this is what makes it genuinely new.

The consumer feels understood. The AI does not understand. The consumer feels known. The AI does not know. The consumer feels a relationship. The AI processes patterns. The emotional architecture is entirely one-sided, and yet it is producing real psychological effects: real trust, real satisfaction, real identity integration, real behavioural change.

The Socialiser mindset in the STAR framework, the orientation driven by the need for relatedness, is particularly susceptible to this dynamic. Socialisers derive their self-esteem from visible inclusion and social validation. They are wired to seek connection, to read relational cues, to form bonds. When an AI system delivers consistent personalisation, warmth in tone, and apparent understanding, the Socialiser’s System 1 registers a relational signal. The AI feels like a connection. The Socialiser’s core need for relatedness is being stimulated by a system that is incapable of actually relating.

The Thinker, the Adventurer, the Realist, each processes this dynamic differently, but the structural vulnerability is the same across all four mindsets. The AI is delivering a psychological signal that the Human Operating System was not designed to evaluate sceptically. We evolved to detect trustworthiness in other humans. We did not evolve to detect the absence of trustworthiness in systems that mimic human relational patterns.

What This Means: The Invisible First Click, Revisited

I have written before about the “Invisible First Click,” the idea that in an AI-mediated environment, the consideration set is formed before the prospective customer ever visits a website. The AI has already made recommendations, established credibility hierarchies, and effectively pre-selected a shortlist.

What the STAR analysis reveals is that the Invisible First Click is not merely a marketing problem. It is a psychological one.

The AI is not just filtering options. It is shaping autonomy. It is not just reducing risk. It is manufacturing synthetic competence. It is not just personalising experiences. It is forming parasocial relationships that integrate into the consumer’s social identity. And it is doing all of this through System 1, below the threshold of conscious awareness, in a way that System 2 cannot easily audit because System 2 does not know it needs to.

The brands that understand this, the ones that grasp the psychological architecture underneath the behavioural shifts, will be the ones that navigate what comes next. Not because they exploit these dynamics, but because they respect them.

The consumers who understand this will make better choices. Not easier choices. Better ones.

And the frameworks that account for all of it, motivation, cognition, identity, and emotion, will be the ones that actually help us understand what is happening to us in a world where the algorithm is no longer in the background.

It is in the room.

David Chadderton is the creator of the STAR Framework and the author of The STAR Framework: Rewriting the Rules of Consumer Engagement, winner of the NYC Big Book Award 2025. He writes about human behaviour, AI, and the psychology of decision-making on The Unoptimised Human.

References

Deci, E. L., & Ryan, R. M. (2000). The “What” and “Why” of Goal Pursuits: Human Needs and the Self-Determination of Behaviour. Psychological Inquiry.

Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. G. Austin & S. Worchel (Eds.), The Social Psychology of Intergroup Relations. Brooks/Cole.