Detecting Fake IDs

Dr. Dawn Weatherford, Assistant Professor of Psychology at Texas A&M University in College Station, Texas is studying the accuracy of detecting fake ID’s in the hope of improving employee training for the transportation security administration, or TSA, and the United States border patrol.

TRANSCRIPT

Dr. Dawn Weatherford, Assistant Professor of Psychology at Texas A&M University in College Station, Texas, is studying the accuracy of detecting fake I.D.s in the hope of improving employee training for the Transportation Security Administration, or TSA, and the United States Border Patrol.

Take a look.

Most of the research that deals with algorithms or training a computer how to recognize faces are moving at incredible rates.

The idea that we can look at our iPhones, and it can detect our information, or they can have drones that look at surveillance and say, this is the person who they think it is, doesn't necessarily speak to all the human observers who have to perform this task on a day-to-day basis.

So cognitive psychology actually can contribute to that process by saying humans will and always probably will perform these types of tasks where they have to authenticate identities in person.

Computers are very expensive.

They'll likely never replace humans exclusively, and even more importantly than that, we have thousands and thousands of people in places of national and international security performing this task right now, and there is no real future for them to be replaced by computers altogether.

So to say that we need to have research that deals with the computer side of things is absolutely important.

We deal with faces all the time.

We, as humans, think we're really good at recognizing faces, so I think it's really important to kind of explore this topic and, you know, are we experts at recognizing faces?

If so, why are we?

What about it?

If not, what can we do to make ourselves better?

Like, who isn't, who is?

So it's really interesting.

Most of what the training entails is, actually, is this card authentic?

So, does this card have the right security strip?

Is this card -- Does it have the right ink?

So the right marks.

And, yet, one of the things that we study in this laboratory is, is that picture the same person as who presented it to you?

It's not uncommon to find that people can switch I.D.s.

So they may have gotten it from a legitimate source, and yet it does not belong to them, and so what my research focuses on is individual ability to recognize the face that's on the card, not just whether or not the I.D. is authentic.

So the typical participant actually sits in front of a computer and is first instructed on how to perform the series of tasks.

So this represents one of our sets of instructions, and just briefly, it tells them, you're going to be presented with two different pictures.

What your task is, is to determine whether or not they represent the same person.

If they do, then you're going to give a high value between 4 and 6 to indicate a high degree of confidence that the answer would be 'yes,' but if you think that they don't, and that would be a mismatch, then you're going to give a low value between 1 and 3.

In this instance, these actually are the same person, if you look closely, although she looks kind of different.

The pictures are taken at least a year apart, so that way it represents a little bit more of what the real world would see in terms of variability.

So if Jasmine makes any choice between 1 and 6.

Here she accepted an authentic I.D.

So she's being given feedback on her correct response.

And if she presses 'spacebar,' she'll be given the next task.

This is an instance of a mismatch.

So if you look closely, you would see that those are actually two different individuals.

Although, in a real-world situation, one could probably pass this I.D. to the other, and someone may or may not notice it.

So we actually compare performance on these types of trials where it's not actually a matching identity with the ones that we saw previously that were matching identities to get a sense of people's accuracy under different sets of circumstances.

I think it's good to know -- like, have science to prove that if we are good at this task, so we know border security and what they're doing is good.

It's also good to know, like, are we doing this wrong?

If so, what can we do to improve or what is -- what are we not doing that needs to be fixed?

I think everyone has something to gain when we focus our attention on the safety of whether you're talking about your city or your state or your country or international kind of considerations, and if this presents a threat, if presenting a fake I.D.

is a way that someone could gain access to a place that they're not supposed to be, and that kind of, perhaps, threatens the safety of everyone, then what I hope to do is contribute even a small piece to making us all a little safer, and here at Texas A&M-San Antonio, ultimately, what it would be great to find is ways to increase security, increase protocols, increase training, and just make everyone a little bit safer.

♪♪