Prediction is now much more an art than a science. Economic forecasting has been shown to be little more accurate than guesswork. Political forecasting is just as suspect: as the political scientist Philip Tetlock showed, experts are terrible at predicting what will actually happen.[1] Worse, many professional futurists continue to repeat old predictions (like the end of work and long-tenure jobs, university campuses and live theatre) even when they've been proven to be wrong.

But my prediction is that predictive tools will move from being of marginal interest to become part of mainstream culture and everyday life.

The main reason is that science is increasingly entering this murky territory.

Weather forecasting has steadily improved - even if only a few days into the future. In business, companies like Wonga have made money through superior algorithms for predicting risk and credit-worthiness (the ethics of the loans that result are of course another matter). When you walk into a GP's surgery the PARR (patients at risk of re-entering hospital) system may be assessing your chances of entering hospital. In criminal justice the Level of Service Inventory-Revised (LSI-R) has been used for a long time, and shown to be fairly accurate. Kaggle runs regular competitions for programmers to develop algorithms which can demonstrate better predictions, for example of school results or health.

In many fields, prediction won't be easy: genetic tests, for example, have been much less useful for predicting disease risks than many expected. But there's no doubting the direction of travel.

Rigour is also being brought to more intuitive kinds of prediction: for example, observing three minutes of a marital row to predict whether the marriage will survive, or the first few seconds of a job interview, or a short stretch of a tutor teaching a class to predict their overall scores.

As predictive models become more mainstream it's possible that our thinking about prediction will change. In the film Minority Report it's presented as a modern version of astrology - if you gaze into the scientific crystal ball you can know for certain what will happen. Instead the opposite is likely. People will become much more aware of probabilities, and therefore more immunised against the idea that the future is knowable. We'll all become more Bayesian - able to understand that the world around us is probabilistic rather than deterministic. That will help us to be better decision-makers. If for example you're given a prediction of your risk of entering hospital in the next two years you may decide to radically change your lifestyle.

Indeed the greatest value from these models lies less in prediction itself but rather in changing our awareness that we can change how the dice fall. The paradox is that where the science fiction of predictive tools tended to diminish the space for human agency, the science may end up expanding it.

[1] Philip Tetlock, Expert Political Judgement. He also showed that Norwegian rats can outperform undergraduates on some predictive tasks.