The Thread You May Have Missed

Four articles. Four different angles. One uncomfortable conclusion that nobody wants to say out loud.

Over the past few months, I’ve written about AI from four distinct directions. I’ve written about trust, specifically why only 11% of consumers will let AI make a purchase decision for them. I’ve written about anxiety, and why Sundar Pichai’s explanation for AI fear is dangerously incomplete. I’ve written about cognition, and whether we’re systematically training humanity into fast, autopilot thinking. And I’ve written about governance, and why the gap between AI enthusiasm and AI rigour is wider than anyone in the boardroom wants to admit.

Each piece stands on its own. Each has its own data, its own argument, its own conclusion. But there is a thread running through all four that I haven’t explicitly named until now, and it’s this: AI is being designed for one type of mind and sold to everyone.

That is not a marketing problem. It is a psychological architecture problem. And it is, I believe, the single most important thing the technology industry refuses to understand.

The Four Symptoms

Let me briefly trace the four arguments, because the pattern only becomes visible when you see them together.

The Trust Problem. Gartner found that only 11% of consumers are willing to let AI make purchase decisions on their behalf. Even for low-risk categories, the number barely moves. Consumers want AI to help them research, compare, and narrow options. But the final call stays human. When I mapped this onto the STAR Framework, the 89% stopped being a monolithic block of “AI sceptics” and became four distinct groups, each resisting for completely different psychological reasons. Thinkers resist because AI lacks transparency. Realists resist because it introduces uncertainty. Socialisers resist because it removes the social process. Adventurers resist because it pre-filters their world.

The Anxiety Problem. Pichai told the New York Times that “humans aren’t evolved to process this much change.” It sounded humble. It was a hand-wave. AI anxiety is not one problem. It is four. Connection-driven people are anxious because AI is replacing the human signal. Evidence-driven people are anxious because AI is opaque. Freedom-driven people are anxious because AI is deciding for them. Stability-driven people are anxious because the foundations are shifting. Pichai’s framing treats all four as a single fever and prescribes paracetamol.

The Cognition Problem. Every major technology company is making the same bet: that you would rather have the AI think than think yourself. Google wants Gemini to be your Universal Agent. Perplexity has eliminated the need to search. Meta AI sits inside your WhatsApp conversations. The cognitive work that used to require slow, deliberate, effortful System 2 processing is being automated. What remains for the human is System 1: recognise the output, evaluate it intuitively, accept or reject it on feel. We are not just adopting AI tools. We are being trained into a cognitive mode.

The Governance Problem. Seventy-eight percent of global companies now use AI. Only 29% report seeing significant ROI. That 42-point gap is not a technology failure. It is an enthusiasm-rigour gap. Organisations are deploying AI at System 1 speed and discovering, at System 2 speed, that deployment without governance is not adoption. It is experimentation masquerading as strategy. The promotion-focused organisations are sprinting. The prevention-focused organisations are stalling. Neither camp is fully right.

The Thread

Four different articles. Four different datasets. Four different conclusions. But one underlying pattern.

Every one of these problems traces back to a single design assumption: that the human on the other end of the AI is a specific type of person, and that everyone else will adapt.

The AI industry is building for the person who values speed over rigour, convenience over control, output over process. The person whose System 1 is dominant, whose regulatory focus is promotion-oriented, whose tolerance for opacity is high. The person who sees a black-box recommendation and thinks, “Great, that saves me time.”

That person exists. They are real. And they are one quarter of the population.

The other three quarters have different psychological architectures. They have different fundamental needs. And the AI being built is not just failing to serve those needs, it is actively threatening them.

The SDT Binding Thread

The theory that connects all four articles is one that psychologists have studied for decades but that the technology industry has largely ignored: Self-Determination Theory.

Developed by Edward Deci and Richard Ryan, SDT identifies four fundamental psychological needs that drive human motivation and wellbeing. Competence: the need to feel effective, to master challenges, to demonstrate skill. Autonomy: the need to feel volitional, to be the author of your own actions. Relatedness: the need to feel connected, to belong, to be seen as a person rather than a data point. Security: the need to feel stable, to have predictable foundations, to trust that the ground will hold.

These are not preferences. They are not personality quirks. They are deep motivational structures that shape how people process information, evaluate risk, make decisions, and experience satisfaction. And every single one of them is being disrupted by AI in a specific, measurable, and largely unacknowledged way.

Competence is being subsidised away. The Thinker’s core need is to feel effective through the rigour of their own analysis. AI doesn’t just provide answers; it bypasses the process that produces the psychological dividend of competence. When the analysis is handed to you pre-formed, the feeling of “I did this well because I thought it through carefully” is replaced by “I accepted this because it looked right.” The Thinker’s competence need is not being met. It is being made redundant.

Autonomy is being quietly eroded. The Adventurer’s core need is to feel volitional, to be the one making the choices. Every AI-mediated recommendation, every pre-filtered feed, every algorithmic curation is a micro-decision that used to be theirs and is now the machine’s. The Adventurer doesn’t feel served. They feel managed. And the difference between a tool and a cage is whether you chose to pick it up.

Relatedness is being simulated. The Socialiser’s core need is to feel genuinely connected to other people. AI chatbots that simulate empathy, recommendation engines that proxy community opinions, and synthetic voices on customer service lines are all attempting to meet the relatedness need without actually providing it. The Socialiser is not anxious about AI because it is fast. They are anxious because it is replacing the human signal with something that looks like connection but is not.

Security is being undermined. The Realist’s core need is to feel stable, to trust that the foundations will hold. AI is shifting the ground beneath every industry, every profession, every institutional trust structure. The Realist is not afraid of innovation. They are afraid of foundations being rebuilt by people who do not value foundations. And the data, from AI-related data breaches to governance failures to hallucination rates, suggests they are right to be worried.

The Design Failure

Here is what makes this a design failure rather than a market failure: the technology industry is not unaware that different people exist. They have the data. They have the user research. They have the segmentation studies.

But the design decisions are being made by a specific subset of people, promotion-focused, System 1-dominant, autonomy-oriented, who are building AI in their own image. They are optimising for the cognitive mode they themselves prefer. They are designing for the user who wants to go faster, who trusts the output, who is happy to delegate the thinking.

This is not malice. It is a form of psychological projection at industrial scale. The builders assume that the user on the other end shares their motivational architecture, because the people they work with, hire, and socialise with largely do. Silicon Valley is not representative of the population it serves. It is a monoculture of a specific motivational profile, and it is building tools that serve that monoculture while assuming universal applicability.

The result is a technology landscape that is spectacularly good at serving one quarter of the population and progressively alienating the other three quarters. Not because the technology is bad, but because the design assumption is wrong.

What This Means

If you are building AI products, the thread changes your design brief. Stop asking “How do we make this faster?” and start asking “Faster for whom, and at what psychological cost?” Stop asking “How do we reduce friction?” and start asking “Whose friction are we reducing, and whose process are we eliminating?” Stop asking “How do we make AI more human?” and start asking “Which humans are we making it more like?”

If you are marketing AI products, the thread changes your messaging. One message will not serve all four motivational profiles. The message that reassures the Thinker (“we show our reasoning”) will bore the Adventurer. The message that excites the Adventurer (“discover something new”) will alarm the Realist. The message that comforts the Realist (“we’ve tested this thoroughly”) will feel controlling to the Adventurer. The message that connects with the Socialiser (“see what people like you think”) will feel manipulative to the Thinker.

If you are leading an organisation that is adopting AI, the thread changes your governance. The 42-point gap between adoption and ROI is not a technology problem. It is a motivational alignment problem. Your promotion-focused leaders are driving adoption. Your prevention-focused leaders are flagging risks. Both are right. Neither is sufficient. The organisations that will succeed are the ones that build governance structures that honour both orientations, that move at System 1 speed with System 2 scrutiny.

The Uncomfortable Conclusion

The thread I’ve been tracing across these four articles leads to a conclusion that is uncomfortable for an industry that believes it is building the future for everyone.

AI is not neutral. It is not universally beneficial. It is not a rising tide that lifts all boats. It is a technology that, by its current design, serves a specific motivational profile and progressively undermines the needs of everyone else.

The Thinker’s competence is being automated. The Adventurer’s autonomy is being pre-empted. The Socialiser’s relatedness is being simulated. The Realist’s security is being destabilised.

And the people building these systems are not asking whether this is acceptable, because the people building these systems are the ones whose needs are being met.

That is the thread you may have missed. And it is, I think, the most important conversation the technology industry is not having.

David Chadderton spent his twenties and thirties teaching people how to make life-or-death decisions at forty thousand feet. He now applies the same principles to consumer psychology, which, depending on the brief, can feel equally high-stakes. He’s the creator of the STAR Framework and the author of The STAR Framework: Rewriting the Rules of Consumer Engagement (NYC Big Book Award 2025), The STAR Operating System: Decode Mindset, Understand Motivation, Transform Human Behaviour, and Dear Algorithm, It’s Not Me, It’s You. By day, a Chief Marketing Officer. By night, a behavioural science obsessive who writes The Unoptimised Human because he can’t stop thinking about why people do what they do.