For all the fashionable rhetoric about educating the whole child, the pinnacle of British schooling is a series of pencil-and-paper tests. This might just be acceptable if these tests didn't seek to sort children, in some senses for the rest of their lives, on the basis of their incredibly limited ambitions.

Everyone - government, teachers, employers - now claims to believe in the importance of lifelong learning. Yet the exam system persists in pretending that intellectual capacity, and quite possibly one's future in the workplace, can be summed up by a few scripts scribbled in the school hall one hot afternoon in early summer. For how long can we go on pretending that introducing more and more tests is the same thing as improving education?

As a result of the absence of serious thought about the role of assessment, the current system (SATs at seven, 11 and 14; GCSEs; AS-levels, A-levels) is trying to do about six jobs at once, some of them at odds with each other. First of all, the Government co-opts the results as an indicator of its own success in delivering education as a public service. David Miliband, the Schools Minister, last week wrote an article welcoming the improved results, the subtext of which was that the Government has evidence that it is doing a good job of educating more people.

The results have, at the same time, been subsumed into the Government's neurotic obsession with target-setting. They set up schools in competition against one another and provide a method of judging their relative performance through the league tables.

Third, and quite differently, exams give shape to the curriculum by dictating what is taught, and by lending continuity and credibility to different subject fields. (For example, much of what we now call media studies might once have been taught in English lessons, but it is now examined, and so taught, as a separate discipline.)

Then there's the role of assessment as a kind of countrywide Hogwarts Sorting Hat. Exams play a part in reinforcing the academic hierarchy by siphoning off students into universities: either top rank, 'new' or none. One of the few genuinely interesting stories to surface in last week's coverage was of research by Paul Black and Dylan William, suggesting that, in truth, A-levels are a poor predictor of performance at university.

And if this is correct - which seems likely, given the exams' interest in discovering what you haven't learnt, rather than what you might - then they are probably even less effective in their role as a form of labour-market currency. In this capacity, they are used to sort students into occupations, roles and status groups. Yet employers claim to want - to take a few random examples - people who can work in teams, who are self-motivated, who know how to learn by themselves. None of these qualities is currently assessed and there are grounds for arguing that as teachers come under pressure to teach to the test, these lessons can fall by the wayside.

Finally, exams are a key source of information for parents and prospective students as they try to make decisions about where to go and what to do next. But here again, they are flawed. GCSEs and A-levels are called qualifications - but for what? It's great to be able to remember dates, and the ability to do so might well be indicative of a broader mental capacity. But you only have to meet a few dons to realise that it doesn't guarantee social or communications skills. It's far from the whole story.

One of the deep pleasures of having children is discovering that they have abilities and aptitudes that you didn't expect - and realising that these are going to stand them in good stead in later life. My daughter is the most networked person I know. Her life is lived communicating and collaborating with her friends. She did her A-levels last year, and did well. But I had a dispiriting sense that the exams were chiefly a hoop through which she had had to jump - largely irrelevant to the qualities for which she will be valued when she starts work. The more it becomes apparent to parents and students that exams are beside the point, the greater is the danger that the edifice will collapse.

The world has moved on, exams have not. Students are assessed as individuals, even though, as the workplace becomes more creative, we are increasingly required to work in groups and spark ideas off each other. Information technology is rarely used for conducting assessment, despite its power to open up new ways of doing things.

Exam boards are under tremendous public pressure to do the job they're already doing more efficiently. Some 1.8 million children a year pass through the formal assessment system, sitting approximately 35 million separate test papers. Mistakes in the setting of questions or in marking are politically disastrous. The next round of exams is never more than six months away, and the numbers involved are growing: twice as many men and five times as many women enrol on undergraduate courses than did 30 years ago. There has therefore been little time to think about whether the exams are relevant to the contemporary world.

I do not suggest that assessment isn't important. But we need a more considered view of its role - and I doubt we'd conclude that the best means of assessment is a high-stakes, snapshot exam that carries the burden of proving every thing from how well David Miliband is doing to whether the girl down the road should go for a job in advertising.

So what might a more appropriate assessment system look like? For a start it would measure a broader range of aptitudes than the current one. It would be like an expanded university admissions form, taking into account collaborative skills, the ability to manage one's own learning, work experience, and the ability to take responsibility for oneself.

Information technology renders this simple. It is possible, for example, to conceive of each individual's owning a portfolio of information about themselves. This might require the student's own input, as assessments at work do; it might involve tests that are sat as and when the student felt ready. It could include the assessments of peers and teachers - which would, incidentally, go some way to restoring a sense of professionalism among teachers, eroded by a sense that they are feeding a giant exam sausage machine.

Of all the competing functions of exams, a modernised system would give priority to offering students the means of deciding what to learn and where to go next, but also offering potential employers a view of skills, competences, achievements and dispositions. By the age of 19, everyone would ideally be clear about what he or she is capable of learning, and have the confidence to go out and do it. Crucially, assessment wouldn't be the end of anything, but part of a process that might continue at different speeds and with varying degrees of effort and enthusiasm throughout life.