Sam Harris, apparently, has recently been defending profiling. Without knowing the detail of what he says and the real context, I thought I'd give my tuppence worth in defending profiling in a basic philosophical way. Here is what I have posted elsewhere:

In really simple terms, profiling should be seen in a consequentialist frame (IF you adhere to moral consequentialism. If not, and you adhere to Kantian Categorical Imperatives, then profiling is likely morally bad in and of itself since people cannot be used as a means to an end. But this would also mean you would tell the truth to the nazi knocking on your door asking where the hidden Jews are (who are in your attic) and commit a number of Jews to death because lying is bad no matter the consequence).

(I am assuming certain things here, such that, for example, we know that Muslims constitute 90% of terrorists in the Western World - this is made up for this argument).

1 - Muslims make up the majority of terrorists.

2 - by checking all Muslim passengers, authorities will identify more terrorists than by checking randomly, or equally across subsections.

3 - by identifying more terrorists, more innocent lives are saved

C - Therefore, it is morally right to check Muslims (by profiling) than to use other methods.

In this argument, moral utilitarianism is assumed.

Points to argue here:

1) - is there a higher proportion of Muslims to non-Muslims in terrorists?

2) - is there any other kind of moral badness in profiling Muslims which outweighs the goodness from saving multiple lives?

For example, I have red hair. IT it could be shown that 99.9% of red-haired people are paedophiles, then I would (though it might piss me off) have no philosophical objection to being checked more than other people with regards to my teaching position. It is common-sense, it seems.

Now, I don't know what Harris argued, but if these premises stand, and if we can establish that the moral good from doing this outweighs the moral bad of the act of profiling Muslims, I cannot see a moral problem with this.

It is more a case of looking at proportion of terrorists who are Muslims, not proportion of Muslims who are terrorists. For example, if 1% of bats in your county had rabies, but negligible other species had rabies, the authorities would target the bats in order to remedy the issue.

However, it is irrelevant as to the fact that there are 10 million rats in your county and that only a tiny proportion of the bats have rabies. The key is that almost 100% of animals with rabies are bats. THAT is the important statistic. It would be crazy to go chasing after rabbits in equal proportion to bats in order to cure the problem of rabies. Also, the relative real number compared to other species is also prevalent. 

IF it can be shown that the process of profiling muslims did more good than bad, would you find it morally good?

Therefore, it comes down to what we can calculate as the benefits (saved lives) against what we can count as negatives (pissed off muslims due to potential human rights violations - though it must be stated that profiling does not necessarily mean human rights violations - this is potentially a straw man). 

The next problem is possibly more unanswerable too. what is the value of 1 (2, 50, 100) lives (saved) versus upset muslims. in other words, would you take 50 saved terrorist victims over pissing off a billion muslims? Is this comparable? Then the second order problems of pissing off muslims creating more possible jihadists. Thus it is indeed a very complex issue.

But certainly on prima facie evidence, if profiling could be shown to inhibit / catch x number of terrorists (and potentially save x number of lives) then it seems to be, on face value, a potentially beneficial system.What do you think? Am I wrong? Of course, the devil is in the detail and it is not to say all profiling is defensible- it depends on the method and execution.