When people talk about banning "profiling" that's what they mean.

A credit score could be made more accurate if you included race and ethnicity, but current models are more accurate than one built simply on race and ethnicity. (This may simply be measuring the recial bias of society, but more data is almost always useful.)

Sure, it may offend people to include race, it's not required. EX: People without children are a higher risk than those with children. On it's own not a big deal, and you can't trust any one thing, but screening even 1 in 3 people is still vastly cheaper than everyone.

In the end we can be just as effective without including race if we simply spend more money to profile more people. Making more people suffer for little gain vs. standing up for your principles is less clear cut than people want to admit. Still, both options are viable.

It's definitely not. My dad had one of those 'extensive questioning' sessions the first time (and only the first time) he flew to Israel because (we think) he is former non-US military. He's also whiter than snow, basically the definition of Caucasian.

You're right that there are many ways to profile, but race and religion have been historically pervasive profiling techniques.

It does seem more fair to simply put everyone through the same process. If you're annoyed at why you (an innocent) have to be subjected to such stringent and time consuming security, that's how innocents who fit "the profile" might feel as well.

If those doing the profiling are interested in being successful, and race and religion prove to be ineffective indicators of threat, then such profiling should stop (or bombs will keep getting through at an inordinate rate). Otherwise, race and religion are simply observations about passengers which have value in directing attention.

Correlation != Causation

If we examine the background of anyone who has ever committed terrorism, every major race and religion would be covered. And so much of crime comes from social context, so at what point could you ever say: it's safe to assume that anyone with X ethnic background is not a threat. You can take any race and put them in a hostile, violent context and see them either continue the violence or die.

Or, if you say the majority of terrorists are men, so let's stop profiling women, I'm sure you would find more women committing terrorism. Then are women the greater threat or is it simply that we've made it harder for men to be terrorists?

I don't see a way to profile without affecting the statistics.

I agree with the gist of everything you've written here. And I think you're slightly missing the point.

> "so much of crime comes from social context"

Agreed. This is why I advocate taking that context into account. Not profiling would mean treating every single individual the exact same way, removing all social context from consideration.

> at what point could you ever say: it's safe to assume that anyone with X ethnic background is not a threat

This is not the point of profiling. The point is to build probability distributions, and hone them.

> I don't see a way to profile without affecting the statistics.

This is literally the point! You want to profile on such an exacting superset of factors (behavior, history, autonomous responses, forensics...) that you cause the incidence rate of terrorism to go down, across the board.

A brittle policy like "only screen middle eastern men" will of course get gamed, but that is terrible profiling. It builds a one-dimensional "profile"--you want a profile that is harder to fool, and based on actual behavioral cues.

> If those doing the profiling are interested in being successful, and race and religion prove to be ineffective indicators of threat, then such profiling should stop

I mean maybe if the profiling is being done by robots, but in reality I have my doubts.

I think humans can do this as well, they just need the right incentive structure (i.e. tied to actual results). CYA security theatre is not it.

I think the bias is too deeply ingrained to disappear without strong, sustained institutional pressure to avoid it.