Note: I'm splitting this off from my earlier post on 23andMe's encouragement of genetic testing of children, since I think this rather speculative argument distracts from the main point of that post.
I mentioned in my previous post that there's a real danger that parents might try to use information from current genetic tests to steer their children in specific directions (and that at least one company is already touting such a test) - but the information available right now from genome scans simply isn't accurate enough to justify such decisions.
For instance, to mention a field close to my own heart, the genes currently known to affect athletic performance likely explain less than 5% of the variation in traits such as muscle strength and cardiorespiratory fitness. Any parent seeking to decide which sport their child would be best suited to would be far better off simply encouraging them to try a whole variety of activities and seeing which ones they were best at (and which ones they enjoyed the most); the addition of currently available genetic data would have marginal benefit, if any.
The same applies to an even greater extent when it comes to traits affecting career decisions - intelligence, personality, even attractiveness. Some of these traits are strongly influenced by genetics, but the actual genetic variants affecting them are currently almost completely unknown, and you should be extremely wary of anyone who tells you otherwise (e.g. these guys).
However, it won't be long before many of the genes underlying many traits relevant to career choices and life success are known, after which the equation is forever changed. Although in the vast majority of cases genetic predictions will still be probabilistic rather than deterministic (since there are many non-genetic factors that affect these traits) it would seem rational to gently steer children towards life paths that are most likely to maximise their overall potential, as determined by both genetic and non-genetic tests.
A society where people make decisions about their children based on genotypes needn't resemble the dismal fatalism of Gattaca. With further education the public will hopefully begin to move away from the deterministic "gene for x" mentality that currently dominates, and towards a more subtle and probabilistic model of genetic predisposition. In addition, it's important to remember that as genetic technology advances so will the rest of medicine, so that by the time we have the ability to predict genetic disease risk we will likely also have the ability to intervene and correct major diseases well before the first symptoms arise.
But even in those situations where medical correction is unavailable (or, more likely, unaffordable), so long as genetic data are used as a guide rather than a sharp constraint on future career options this technology will generally have positive rather than sinister implications for society.
To further the athletic example above: rather than using genetic information to block children from becoming athletes altogether, coaches and talent scouts might use it to guide promising young athletes into the sports to which they are best suited. The same model could apply to any other arena. So long as it can be shown that genetic predictions are both accurate and relevant, I find it hard to see how it could be unethical to use them to help kids make choices about their futures; the kids themselves will still be the only ones who can determine whether or not they will succeed. Naturally, careful industry regulation will be required to ensure that this model - rather than some naively deterministic algorithm - is the future of gene-based career counselling.
Please note that I don't believe that genetic testing will ever be used as the be-all and end-all of career guidance, but rather as a tool for career counsellors and other professionals (e.g. athletics coaches) to add to their existing battery of tests. It will likely turn out that for many traits the genetic data is already effectively captured by simply assaying direct or indirect measures of the trait itself, making genetic tests completely redundant in such cases. However for other traits - particularly where there is a need to make early predictions before the trait itself is evident - genetic tests may provide a useful tool.
This is all pretty speculative right now, but these are issues that society will have to wrestle with very soon, probably much sooner than we are really ready for them. It's obvious that DNA-based trait prediction tests will be popular with many parents, as evidenced by the fact that companies are already offering them despite the laughably poor predictive power of currently known genes. The sooner we start having a constructive and realistic discussion about the best way to use these tests, the better.
