A reflection on why prediction, once acted upon, stops being a neutral observation and becomes an intervention in the life of the person being predicted.
Human beings behave differently when they know they are being evaluated. An employee works differently when a manager is observing. A student prepares differently when the assessment criteria are known. A public speaker becomes more careful when every sentence is being recorded.
Artificial intelligence adds another layer to this behaviour. Increasingly, systems do not only evaluate what people have done. They predict what people are likely to do next — who is likely to succeed, leave an organisation, repay a loan, develop a health problem, respond to an offer or create future risk.
Prediction can support better decisions. An organisation may identify an employee who needs support. A healthcare provider may intervene earlier. A business may understand customer needs more accurately. But predictions do not remain outside the lives of the people being predicted. Once acted upon, they can alter the environment around that person.
An employee classified as unlikely to remain may receive fewer development opportunities. A customer identified as low-value may receive poorer service. A student predicted to struggle may be offered simpler work and fewer chances to demonstrate unexpected ability. The prediction then begins to influence the outcome. The person may eventually behave in a way that appears to confirm what the system believed from the beginning.
This is more than algorithmic bias. It is a behavioural loop between the prediction, the institution and the individual.
My earlier research compared how cultural characteristics were expressed in Thai society and in virtual communities. Its broader implication was that behaviour cannot be treated as a fixed personal characteristic. The surrounding environment influences which behaviours become easier, safer or more rewarding.
AI prediction may overlook this relationship. A model sees a pattern in historical behaviour and converts it into an expectation about the future. But the future behaviour may depend partly on how people are treated after the prediction is made. This creates a difficult question: Is the system recognising a person’s likely future — or helping to produce it?
People may also change themselves when predictions become visible. An employee who learns that a system has classified them as having limited leadership potential may lower their ambition or begin performing the style of leadership the model appears to reward. A creator may adjust their work toward what an algorithm predicts will succeed. A job applicant may organise their language, experience and personality around what automated screening is believed to recognise.
In each case, the person becomes partly shaped by an anticipated machine judgement. Over time, this may narrow the range of acceptable human behaviour — people who fit the recognised pattern receive opportunity; people whose potential appears in unfamiliar forms learn to translate themselves into something the system can predict.
The intelligent age therefore requires us to treat prediction as an intervention, not simply an observation. Organisations should examine what changes after a prediction is introduced. Does the classification increase support or quietly reduce opportunity? Can a person challenge the conclusion? Can the prediction expire? Can new behaviour outweigh old data?
Most importantly, do decision-makers understand that probability is not destiny? AI can reveal patterns across large populations. Human beings, however, do not live only as members of statistical groups. They learn, recover, mature, respond to encouragement and sometimes behave in ways their past would not have predicted.
Good systems should help institutions recognise risk without closing the future. A prediction should create a question for human judgement — not a script the individual is forced to perform.
The moment a prediction changes how we treat someone, it stops merely describing the future and begins participating in its creation.