Why More Data Doesn't Mean Better AI Decision Making — The Deficit Model Explained
- David Harvey
- 6 days ago
- 5 min read
There is an assumption buried inside almost every AI decision tool ever built.
It is rarely stated explicitly. It does not need to be. It is so deeply embedded in the way these tools are designed that it operates as a silent premise beneath everything they do.
The assumption is that people make bad decisions because they lack sufficient information.
Give them more data. Better analysis. Clearer charts. More options. More confidence percentages. More scenario modelling. And they will decide better.
This assumption has a name. Professor Hannah Fry and Michael Stevens call it the Deficit Model. And the evidence against it is overwhelming.
What the Deficit Model gets wrong about AI decision making
The Deficit Model emerged from science communication research. It was the dominant framework for decades if the public resisted expert recommendations on vaccines, climate, or public health policy. The explanation was simple: they did not understand the science. Give them better information, and they would come around.
The problem is that it does not work. People who receive more information do not reliably change their decisions. Sometimes they entrench further. The data deficit was not actually the problem.
The real barriers to better decisions are structural, emotional, relational, and contextual. They are about how information is framed, not how much of it exists. They are about the emotional state of the person making the call. They are about trust — in the source, in the process, in themselves.
Hannah Fry puts it precisely: the person with the most data does not make the best decisions. The PhD and the person without one both decide emotionally. The difference is not the quantity of information available. It is the structure they apply.
Michael Stevens arrives at the same conclusion from a different direction. His work in participatory learning and science communication makes the same argument: people do not need more access to information. They need structures that help them make meaning from the information they already have, in their own context, on their own terms.
Information access was the twentieth-century problem. Decision structure is the twenty-first-century problem.
Why most AI decision making tools repeat the same mistake
Look at how most AI tools are positioned, and you will see the Deficit Model running underneath all of them.
More outputs. More bullet points. More analysis. More data sources integrated. More confidence scores. More options generated.
All of it assumes that the reason the decision is hard is that something is missing from the information stack. Add the missing piece, and the decision becomes clear.
But most hard decisions are not hard because information is missing. They are hard because the framing is wrong, the person's emotional state is unresolved, the relevant factors have not been separated from the irrelevant ones, or the decision has been running in a loop long enough that the person can no longer see it clearly.
More data does not fix any of those problems. In many cases, it makes them worse. Analysis paralysis is not a shortage of information. It is an excess of unstructured information with no governing framework to tell you what matters and what does not.
What Ask Omega* does instead — governed AI decision making
Ask Omega* was built on the explicit rejection of the Deficit Model.
The question readiness scorer does not provide additional data. It helps you clarify what decision you are actually trying to make — because most looping decisions are not information problems, they are framing problems. The blocker is not missing evidence. It is an ungrounded question.
The Lens system does not add more analysis. It translates the same decision through different cognitive and emotional filters — Wife 2.0 for plain truth, The Clinician for evidence and risk, The Gamer for optimal path and failure states, The Elder for long-arc perspective. Because the barrier to action is not missing data. It is missing the right perspective on the data you already have.
The Comfort Index does not measure information quality. It measures whether clarity arrived. Whether the decision stopped looping. Whether the felt sense of resolution came. Because the goal of a governed AI decision making system is not the accumulation of evidence — it is the moment the person feels clear enough to act.
And the Wife 2.0 lens exists for precisely the reason Fry identifies: the Deficit Model version of advice — here are the facts, here is the analysis — often fails to produce action. What stops the loop is plain truth delivered in a voice the person trusts. Not additional information.
The signal that proves the Deficit Model wrong
After every decision receipt in Ask Omega*, we surface a single binary prompt.
Did this help? "Spin stopped". or "Not yet".
That prompt is not a satisfaction rating. It is not a data quality score. It is a test of whether the felt resolution of the decision loop arrived, which is exactly what the Deficit Model cannot measure and does not try to.
The neuroscience underneath this goes back to Antonio Damasio's somatic marker hypothesis. The research shows that patients with damage to the emotional centres of the brain become paralysed by decisions despite perfectly intact rational reasoning. Emotion is not noise in the decision process. It is the signal that tells you which option matters. Without it, analysis produces infinite loops.
The Comfort Index is the product expression of that insight. Not a replacement for the emotional decision — an environment in which the emotional decision can be made from clarity rather than from anxiety, incomplete framing, or a question that has been looping long enough to lose its shape.
The advantage that actually matters
The advantage in an AI-driven world will not go to the person with the most data.
It will not go to the organisation with the most powerful model, the fastest output, or the most impressive benchmark score.
It will go to the person with the best structure for the specific decision they are actually facing — right now, with the information they already have, in the emotional and contextual state they are actually in.
That is what governed AI decision making looks like when it is built on the right premise.
Not more data. A better decision.
Ask Omega* is available now for founding Pro Users. The 7-Day Comfort Index Challenge is the entry point — seven governed decisions, one per day, with a visible record of how your Comfort Index moves when the structure around a decision changes.
Comments