Faiza Patel: Our thought today was to maybe have a bit of a discussion about what we actually know about extreme vetting, what it’s about. What are its risks? What are its benefits? Do we need it? Do we already have it? Those kinds of things.
I’m joined today with two people who are very, very expert in these issues. David Bier from the Cato Institute is an analyst on immigration issues and has published extensively on these issues, including some really excellent statistical analyses that delve into how and when immigrants have, the very rare occasions in which immigrants have committed terrorist acts in the United States, but also looking more closely at how vetting might impact our overall system of people coming into the United States.
To his right is Alvaro Bedoya of Georgetown Law School who has been taking the lead in looking at the extreme vetting policies that have been proposed by ICE, which would impact not just people applying for visas, but potentially people who are in the country with visas, potentially also permanent residents of the United States.
These are experts in their own fields. What I thought what I would do to start out is to sort of tell you what we found as we started to dig into these issues and where our areas of concern are.
I think from our perspective, we know that this administration has a stated policy goal of repeated banning Muslims from the United States. Having looked at the extreme vetting initiatives that have come forward, it seems pretty clear that this is simply another way, a backdoor way, of implementing some of those policies. We anticipate that it will and already has generally dampened travel to the United States and have a very high cost for our economy, for our values, and also for what we are as a country.
With that top line framing, I want to talk a little bit about what’s been happening on the ground. Prior to a DHS study, which was put forward as part of the justification for the most recent presidential proclamation on travel from certain countries, there was very little evidence to suggest that our current visa vetting system required an absolute and complete overhaul. I don’t think anyone takes the view that the vetting system is perfect, and I’m sure there are lots of ways in which it could be improved, but there’s very little to suggest that a whole new approach to visa vetting was called for.
In fact, the visa issuing system of the United States, I would say, has long been recognized as being one of the most strict in the world, and a big concern from businesses and immigration lawyers has actually been that it’s too rigid and it’s too difficult for people to get through the system rather than that it’s just too easy and people are coming in willy-nilly.
A fundamental principle of that system, which I think is really important to keep in mind as we look at these different measures, is that the burden in the visa’s application process is on the applicant. Whoever wants to come to the United States, whether it’s for business, whether it’s for sightseeing, or for education, there’s a burden of convincing the consular officer that he or she is eligible to come to the United States. So, if there’s insufficient information available to a particular individual in a particular country that would allow them to convince the consular officer, they just don’t get a visa.
The other thing I think is really important to remember about the visa vetting process is that at least since 9/11 we have built in very robust national security checks. There are a number of databases, some of which will flag people automatically, others which will require a consular officer to probe deeper, but we have a very multi-layered system of checking people who want to come to this country.
I guess the proof is in the pudding, because, despite the hype from some corners, the likelihood of someone getting killed in a terrorist attack by an immigrant is one in?
David Bier: As of November 2017, it’s one in 3.8 million per year.
Faiza: That’s a pretty significantly tiny percentage.
Now, getting down to what’s going on here, right? You have this framework. You know that the system has been in place and has been effective in keeping our country safe.
Well, I think that part of the issue here is the lack of transparency. We know that the administration at least claims to be doing something called extreme vetting, but now it’s a question of what is extreme vetting exactly? We have some clues about what it is and or what it might be, but we don’t actually have the full picture. So, what we might tell you today is necessarily limited by the information that we have available.
One thing that has caused some alarm bells to go off is the idea that the administration intends, based on State Department rulemaking, to identify groups of people who they think warrant further scrutiny. This goes back to the point I was making earlier about the individual applicants having the burden of demonstrating that they actually are entitled to a visa, not entitled, but that they should be granted to a visa.
There’s a lot of concern on our part that this group vetting mentality is starting to take hold in the vetting system. We have a great deal of concern that this is going to be used to hold up travel from the countries, such as those that were affected by the three iterations of the Muslim ban orders.
One reason that we think that is that the number of people that the State Department estimates would be affected by these new procedures, so it happens to coincide with the annual number of visas from those countries at large. That’s around 65,000. I think it’s up to around 70 this year. But, it’s really quite close to that. It’s a really striking fact.
We don’t actually know what criteria the State Department would use in defining populations warranting additional scrutiny.
I want to be clear over here that when we say additional scrutiny, that could very well mean an absolute bar on these people coming to the United States. As it is, there’s an extensive amount of documentation that any individual who seeks to travel to the United States has to provide to consular officers, and increasing those burdens more and more will effectively stop people from coming into the country. Anecdotally, we’ve already been hearing from people who work on these issues on the ground and from immigration lawyers that people are just becoming more and more reluctant to go through the process and to go through all of the hoops that are required to come to this country.
I think that the other concern we have is that given the anti Muslim crucible in which these policies have been born is that there’s a considerable risk that they’re going to devolve into a form of ideological vetting. This country has engaged in ideological vetting in the past. It’s something that we’ve moved away from, that Congress has moved away from, and it’s an approach that’s been rejected by consular officials who have come forward to say it’s not really that effective in identifying real threats or dangers to public safety.
Also, I think for me at least, somebody who immigrated to this country, as well, I think about the fact that I was ideologically profiled and people were looking at what my views were, I’m not sure that I would pass that test. I think it’s a very dangerous path to go down.
It also, to me, really cuts against the grain of the United States as a country, which is devoted to the free and open expression and exchange of ideas. That’s the fundamental premise of our democracy. The idea that we’re going to be excluding people not because they present a particular danger, but because we don’t like what they say, I think is really, really troubling.
I think concerns about ideological vetting for me are deepened by the emphasis in the current system on social media. We all live in this world where social media seems to be everything every day. There seems to be this assumption that we can use social media to glean all kinds of information about people.
I think there’s several concerns around this emphasis on social media. There’s three major ones that strike me. One is obviously this relationship to ideological monitoring, particularly because some of the standards that have been suggested in the social media monitoring context, such as whether or not somebody will contribute to American society, are very vague and very malleable, and allow all kinds of disproportionate targeting of particular faiths or ideologies.
Free speech. There’s no question that knowing that being watched or that what you are saying online is being monitored will naturally make you much less willing to put your actual views out there. I was thinking about this.
I have a cousin who lives in Saudi Arabia. I’m on WhatsApp with her and I chat with her back and forth. Recently, then all of these things happened, happening, in Saudi Arabia. I would’ve loved to have sent her a message and say, “Hey, what’s going on over there? How are people reacting to this?” But I didn’t, because I was afraid that what I was going to say to her and her response to me was going to be seen by somebody, and I don’t want to get her into trouble, because she lives there.
It’s a very real thing that people go through, and I think it’s important to understand that the chilling effect is not just for people who are overseas, but also for the people who they speak to in the United States. Whether they be family, whether they be friends, whether they be business associates, there’s a whole range of people here whose speech gets chilled because of what we’re doing in terms of the visa vetting process.
The last thing I think that’s worth mentioning in terms of the social media issue is that it’s really unlikely to make us safer. If anybody wants to do harm to the United States and they’re going to the visa process, the first thing they’re going to do is scrub their social media. They’re not going to be leaving, “I heart ISIS,” messages on their Facebook page. That’s going to be gone.
The second reason, and I think this is one that’s really, and I think Alvaro will talk about this a little bit, is that social media is uniquely open to misinterpretation. Oftentimes, it’s short form. People use symbols. It’s kind of hard to know what it means. I might like something on Twitter not because I like it but because I want to remember to go back and read it. There’s all kinds of ambiguities around social media, which are very easy to misinterpret.
But, even beyond that, I think, the idea that we can actually predict what someone is going to do based on social media, so we can predict dangerousness, or we can predict threat to public safety, is a very far stretch. Particularly given the volume of immigration to the United States and all these difficulties around what social media means, your error rate in that is going to be really high, which has two consequences. One is that your consular officials are going to be off chasing red herrings, and the other one is that people who have a legitimate reason to come to the United States will be tagged unfairly and wrongly as being a threat. All of these things contribute to making social media not a very effective way of making us safer.
To that end, seems worth noting that a lot of these measures are being implemented behind closed doors, they’re not under review by any court, and their impact is going to be very difficult to document and to assess. This is going to be dispersed over lots of consular offices around the world. Individuals getting their applications denied, we don’t know really know whether the reason for that is because of some discriminatory program or policy or whether it’s a regular visa denial. It’s really hard to get data on that kind of implementation.
This is one of the reasons we wanted to come here and talk to you guys, because we think it’s really important that Congress, which has authority over our immigration laws, and has designed these laws, and has put forward its priorities as to what needs to be emphasized in laws, actually takes the steps necessary to push for transparency from the executive branch, the State Department, and the Department of Homeland Security in particular as to what is actually going on and how these policies are being implemented.
I’m going to stop there and turn it over to David.
David: Hey there.
Oh, yes, the microphone.
As was mentioned, my name is David Bier. I am one of two immigration policy analysts at the Cato Institute. The Cato Institute is the leading Libertarian think tank in the United States, perhaps the world. Really, our principles are free markets, individual liberty, and peace.
We believe in the right of Americans to associate themselves with people born in other countries. Really, this presumption that we do not have the right as Americans to associate ourselves, interact with, contract with, in any form or fashion, with people who happen to be born in other countries really cuts against those principles of our personal liberty, as well as the ability of the market to operate freely, and really, the fundamental principles of our society.
But, what I really want to talk about today is the travel ban and the extreme vetting in its own terms. I think if we look at the travel ban, it doesn’t just say that we’re going to ban certain nationalities. It purports to tell us why it’s doing what it’s doing. This is really an insight into how the administration is deciding who is a threat, who is a concern to them.
What the travel ban says is that the Department of Homeland Security created nine criteria that every government in the world was required to meet. At the end of a review process, it banned eight nationalities based on their failure to meet this set of criteria that the Department of Homeland Security came up with.
This did not happen, okay? Know that this did not happen in part because the proclamation tells us that Iraq failed the criteria, but was left off the list. But, more significantly, if we look at the government data on these criteria, we discovered very quickly that many, dozens, of countries around the world fail these criteria that the government itself deems as risks, and yet were left off the ban.
Take one criteria for example, the requirement that the government issue an electronic passport for its nationals for travel to the United States. 86 countries do not issue, in 2017, did not issue an electronic passport to their nationals at all, and many other countries did not issue, had their nationals operating under old passports that did not have an electronic component to it. Yet, only five of those 86 are on the travel ban list.
Another criteria is that they’re not a terrorist safe haven. Well, there are 13 terrorist safe havens according to the State Department, but only four of those 13 are on the travel ban list. You just go down the line.
Another criteria … Another criterion is regularly refusing to accept people that the U.S. is trying to deport. According to Immigration and Customs Enforcement, in 2017, there were 12 countries that regularly refused to accept people that the U.S. was trying to deport. Turns out, only one was on the travel ban list.
So, not only are there many countries around the world that are failing this criteria, which are not on the list, but many of the countries that are on the list are also not failing the criteria.
So, what is going on?
I entered all this information into Excel and loaded it up in a report trying to find any kind of pattern. Is there any group of criteria that would predict an inclusion of these eight countries? There’s no set of criteria that could explain, in the government’s own data, in the government’s own criteria, this set of countries. So, there’s some other reason why these countries are on the list. We don’t know what those are, but it’s certainly not these criteria.
The other thing is if you look at the travel ban, they say that whatever the reasons why these countries ended up on the list, their failings have resulted in less reliable vetting for these nationals, which means that they are more likely to have terrorists and other criminals entering the country from these places, for whatever reason.
One is the process, right? They don’t issue electronic passports supposedly, or they don’t do these certain things.
But, then there’s the outcomes, and they’re saying the outcomes are worse in these places. This is, again, a testable claim. We can look at where terrorists come from.
One of the things that I did was I put together a list of every person who entered the United States for the first time since 9/11 and who was later convicted of a terrorist attack or killed while committing a terrorist attack and identified their nationality. Turns out that had the president’s ban been in effect continuously since 9/11, not a single terrorist who planned, attempted, or carried out a terrorist attack on U.S. soil would’ve been prevented from entering the country. That shows how low the risk is from these nationals who are vetting visas to come to the United States.
My colleague, Alex Nowrasteh, went back even further looking at every person who killed anyone in a terrorist attack, regardless of how they came here, visa, refugee, whatever, over the last 43 years from 1975 until November 2017, and concluded that no national of these countries killed anyone in a terrorist attack on U.S. soil during that entire period of time.
Alex also looked at criminality, generally, reviewing the Census Bureau statistics on incarcerated persons in the United States. Turns out U.S. born Americans are more than five times as likely to be incarcerated in the United States as nationals from these countries. This is controlling for population. So, they have a much smaller population of these eight nationalities, but if you look at them, they are five times less likely to be incarcerated.
Just to conclude with the big picture, as was mentioned, the annual risk of being killed by any foreign terrorist on U.S. soil from 1975 until November 2017 was one in 3.8 million. For comparison, the annual risk of death in a regular homicide, non terrorist, was one 14,000 annually over that same period. We’re talking about orders of magnitude difference between the threat of foreign born terrorism and the threat of just regular homicide, and yet, we’re promoting this picture of terrorism, this existential threat to the United States.
What about vetting? Someone could get into the United States who’s a terrorist who then doesn’t kill someone. Since 9/11, when the vetting system was totally revamped, only 27 foreigners entered the United States as adults or older teenagers over the age of 15 and then went on to commit a terrorist attack or offense of any kind, including supporting terrorism overseas, including traveling to overseas. These are people who are not necessarily trying to attack the United States, only 27, so less than two per year.
This is the potential pool for vetting failures. We’re talking about 27 people. That’s less … Of those 27, if you look at their history and their background, turns out at most, only 13 radicalized before they entered the United States. However you want to define that, as broad a definition as you possibly can come up with, only 13. Those are your vetting failures. These are the people who may have had an intention coming to the United States. That’s assuming that basically anyone who we don’t have information about was a vetting failure.
This means that from 2002 until 2016, 29 million visas or entrants without a visa were approved for every radicalized future terrorism offender that was approved, 29 million visas or entrants without a visa. That’s an enormous number. Only one of these 13 people actually killed anyone. It was a Pakistani national who shot 14 people and their husband in San Bernardino. Thus, the rate for deadly attackers who were approved was one in 373 million vetting approvals over this entire period of time. That is just an astronomical … The haystack is so huge it’s just almost impossible to imagine that the State Department could be doing a better job than it is.
This won’t fix visa vetting, because based on all the numbers that we have, visa vetting just was not broken before it was issued. Thank you.
Alvaro Bedoya: I want to start by saying if you haven’t read David and Alex’s study at Cato on these probabilities, you really should. It is sobering. It’s a short document like 10 pages, and the executive summary is really powerful.
With that said, let me … I’m going to talk about something that most folks haven’t talked about yet. You’ve heard about the travel ban. You’ve heard about the Muslim ban. I want to talk about the digital Muslim ban, which is something that’s getting a lot less attention.
What am I talking about? January 2017, President Trump issues his first travel ban, the first Muslim ban. Most people rightly focus on these really broad country laws saying, “You’re from these countries. You can’t come in. You can’t come in the United States.”
But, there’s a separate section of that first ban, Section IV, which said something else or additional. It’s a really [inaudible 00:25:39] system to screen everyone coming into the country to figure out if they intend to commit a criminal terrorist act, if they are going to forego the national interest, and if they’re going to contribute to American society.
If you’re listening to this and you’re a little confused, you’re right, because this provision baffled experts, because number one, as David and Faiza have explained in a very detailed way, we vet people very aggressively already.
And then these other criteria, will you contribute to the national interest? Will you further American society? These are made up out of whole cloth. Those don’t exist anywhere in American law. These are totally new, brand new criteria that give ICE and DHS maximum leeway.
But, what happens when the president issues that travel ban? There are legal challenges, courts stop the ban from going into effect, and the president himself withdraws the travel ban. Later bans that come into place only have the criteria about vetting people for crimes and terrorist acts, so that other broad criteria really go away.
What happens? That was January 2017 the order is rescinded.
July 2017, ICE, Immigration Customs Enforcement, posts procurement documents online for something they call the Extreme Vetting Initiative. Suddenly, they do want to screen for whether you’re going to contribute to American society, whether you’re going to further the national interest, and whether you intend to commit a criminal or terrorist act. The exact same criteria that we all thought went away in January resurfaces in July in this screening system that ICE wants to hire a company to build.
How on Earth are they going to decide whether you will contribute to American society, further the national interest, or intend to commit a criminal or terrorist act? They’ve been trying to make these determinations as much as possible automatically through a couple different techniques. Previously, this was a manual process. They want to do this now automatically from the moment you apply for a visa, past the moment you enter the country, and while you are in the country.
So it goes automatically on an ongoing basis, and what do you look for? They’re going to scan media, and these of course they’re going to scan; Facebook, Twitter, LinkedIn, blogs, academic websites, hearings, conferences, everything. They’re going to do social media analysis, the company they’re going to hire is going to do this, do social media analysis to automatically flag people for deportation and visa denial based on whether they have contributed to the national interest, furthered American society, etc.
Who are they going to do this to and how many people are they going to catch up? Well, the contract documents expressly say, “We want you to flag a minimum of 10,000 people a year under these criteria and some of these techniques.”
If you’re wondering, “Okay look, I am probably okay vetting people who are abroad and want to come into the country,” that’s not what this is about. That’s not the only thing this is about, because vetting continues throughout the entire process. The documents repeatedly hint that they’re going to scan the social media of not just visa holders, but also permanent residents.
Then, finally, think about the Internet. There is no American Facebook, foreign national Facebook. There is no American Twitter and foreign national Twitter. There is a Facebook, there is a Twitter, and if you interact with anyone, American citizen or not, who fits one of these criteria, your information will be scanned also.
This is an extraordinarily broad program coming down the pike, and you should be concerned about it, and your bosses should be concerned about it.
Let’s drill on what the problems are. This is all about three problems.
If you want a concise summary of this, I can point you two letters, which are available in an article in Associated Press and Reuters. One is from 50 plus civil society organizations, everyone from Committee to Protect Journalists, ACLU, NAACP, the Pan American Foundation, all groups saying this is wrong.
Another is probably more interesting. It is from 50 plus of the nation’s leading scientists, folks at Google; folks at Microsoft; folks at MIT, Carnegie Mellon, Princeton, Stanford, blah, blah, blah. Literally, the top nation’s scientists saying, “This program will not work. Say what you will about the system, you cannot build this. Instead, this is going to be a high priced, high tech boondoggle.”
One of the three reasons I think you should be concerned about this and your bosses should be concerned about this, the first is free speech. We see extraordinarily broad criteria of who they’re looking for, and with these automatic flags going up, are you going to post an article online critical of someone in the federal government if you think that ICE is watching and they automatically flag you for visa denial or deportation? Are you doing to do that? I think the answer for a lot of people is going to be no.
Again, this is not just about visa holders. This is not just about people who are abroad. This is about people who are in the United States, who are friends and family of American citizens. This is everyone, and this is a real free speech issue.
The other point is that this is a tool for discrimination. There’s a reason this is called the Extreme Vetting Initiative, or it was called the Extreme Vetting Initiative. I’ll explain in a second. There’s a reason this is called the Extreme Vetting Initiative, there’s a reason that this gives maximum leeway to ICE, and there’s a reason the criteria for the system are pulled directly from the original Muslim ban, and it’s that this is not a high tech tool to catch bad guys. This is a high tech smoke screen that will empower ICE to flag whoever it wants, whenever it wants, for deportation or visa denial.
Finally, even if you are more skeptical of immigration and immigrants, even if your boss is more skeptical of immigration and immigrants, and maybe what you primarily care about is the government not wasting taxpayer money, then I think it’s really important to consider that this system cannot and will not work. Again, do not take my word for it. Take it from senior artificial intelligence experts at Google, Stuart Russell, Raj [Ka-men-pa-tee 00:32:40], basically the world’s leading authorities on automated decision making and AI who say, “You can’t build this system.”
Why, I keep on saying that word, why can’t you build the system? As Faiza was alluding to and as we all know, social media is really complex. If you don’t belong to a particular social group, if you’re … When I read my niece’s tweets, I can’t make, and I’m not trying to be cute, I really don’t understand them half the time.
Additionally, we’re asking computers to do this. It’s not like these are unanswered questions. These are questions that have been asked, and the most sophisticated social media analysis systems struggle to make very basic determinations of is this a good tweet or is this a bad tweet? This is proven, and yet we’re asking these systems to decide will this person favor the national interest? Will this person contribute to American society? These scientists are saying you cannot build a system to do that.
The other has to do with probability. Terrorist attacks, when they occur, are horrific and they hurt a lot of people, but as a mathematical fact, terrorist incidents in the United States are too rare for modern predictive systems to accurately predict. Predictive systems are good at predicting things that happen with some regularity, a home run, the stock market, things like that, these systems hitting a certain level. These systems fail when they’re asked to predict things that are mathematically extremely rare.
Finally, these systems fail when they’re asked to make judgments about highly complex, undefined criteria, and the criteria that we’ve been talking about are precisely those things.
So finally, if you have any doubt in your mind that this system that ICE wants to build is about quantity and not quality, consider the fact again that ICE is asking the contractor to build this a minimum of 10,000 individuals a year based on this methodology. Whatever contractor is tasked with doing this is going to have to flag 10,000 people no matter what and no matter the quality of those investigative leads.
What I think you should do, briefly in closing, what I think your bosses should do, I think they should ask questions. I think they should ask questions like, “Why are the criteria for the system pulled from a rescinded executive order, one that literally does not exist in American law today, in any capacity, even an executive order? Why is ICE signaling that it’s going to scan green card holders? How is ICE going to scan foreign nationals’ social media and not scan the media of pretty much every American? And how is this thing going to work if the nation’s leading scientists say it cannot work? And how is it worth taxpayer money to build that kind of system?” That’s something that I think every one of your bosses is in an excellent position to do, to ask questions, conduct oversight, and figure out why this system is being built.
The last thing on this, is some folks at the ICE and DHS have realized that this branding of Extreme Vetting Initiative is probably not strategic, so they changed it to the Visa Lifecycle Vetting Initiative. They’re saying, “No, it’s really early to make judgments about this.”
But, what you need to know is that these documents also say that ICE wants to roll out the system by September of 2018, which in the world of government contracting is right around the corner, so this system is not that far away. You can change the name, but the system is still there. And I think folks need to play a lot of attention to it.
Faiza: Great, thank you so much, both of you. I thought I would just ask a couple more, some follow up questions if I can, just to get a few things out.
To start with you, Alvaro, one of the things that always comes up in this context is, “Well yeah, you’re right. This whole social media monitoring thing is kind of creepy, but we put all this stuff out there. We post on Facebook. We post on Twitter. We post, some of us, on LinkedIn, as well.” So, given that all of this information is already out there in the public, how can we suggest that the government not take advantage of this resource and that they not mine this information for whatever insights it might be able to give us?
Alvaro: Right. I’d say two things to this. The first is this is really 1960s thinking about privacy. Historically, yes, what you have volunteered to the public you haven’t had a constitutionally cognizable privacy interest in. But, this whole idea that what you volunteer to the public you don’t have a privacy interest in, that idea was developed in a world where we didn’t have computers that can analyze literally everything everyone says automatically in a fraction of a second; in a world where computers didn’t exist that could literally track your face down the street and identify you by name, remotely, and in secret and far away, and at significant scope; where there wasn’t technology that could pull down the cell site signals that your cell phone generates and locate you and all your friends and family, again, automatically, at scale, and at cents on the dollar. So yeah, in the past, we didn’t have a privacy interest in what we did in public, but that’s generally because the government would only track people comprehensively, and would tail people with officers, would ask someone to look for you in someone’s writing, if you were up to no good. Nowadays, the government has the ability to scan, look for, and track everyone, again, at cents on the dollar. The other issue is that this isn’t just about privacy. This is about speech, and if what I post on my Facebook profile, what I post on Twitter, what I post on an academic website is not speech, I don’t know what is. When the founders wrote the Bill of Rights, they didn’t say, “American citizens have the right to free speech,” right? They say, “The people have right to free speech.” If you’re living in this country, you have a right to free speech. I do know, even if you disagree with me, there is no foreign international internet and American internet, there’s the Internet. If ICE wants to scan it to automatically flag people for deportation, they’re necessarily going to be scanning the communications and speech of every American essentially.
Faiza: Just to follow up with that a little bit, do you see any tension between arguing on the one hand that social media is an incredibly large pool of information that you can learn a lot about all of us by analyzing these publicly available streams of data, and at the same time saying, “Yeah, it tells you a lot, but it’s not going to help you counter terrorism?” How do you square that [inaudible 00:39:49]?
Alvaro: Actually, I think you’re going to scare a lot of people into not posting their political views online. If you look at … There was a study from two years ago that looked at-
Faiza: Wait, wait, but how’s there a privacy interest? If you’re not revealing information about us, how is that a privacy concern?
Alvaro: I do think there’s a privacy interest, and I’m happy to talk about it at length, but you’ll notice the three things I talked about were not, “It’s going to violate your privacy and it’s going to …”
People are going to censor themselves online, it’s going to discriminate against people, and it’s going to literally not work. We’re going to spend millions upon millions of dollars building something that doesn’t work.
I’m happy to talk at length about my academic notions as to why it violates your privacy, but it’s going to violate your free speech. I think that’s something probably, as a privacy advocate I’ve been saying, but probably a lot of people are going to get it more readily than the privacy question.
Faiza: Don’t you think that some of that speech might help in identifying threats?
Alvaro: Sure, okay. It would be one thing if DHS, and they do do this to a certain degree. I don’t want you to leave this briefing room thinking that DHS has never scanned social media and now they want to scan social media. That’s not true. It’s just a lot more tailored in the past.
It would be one thing if ICE was proposing a system, which does work and is commercially viable, which is to say, “We know what videos ISIS posts. We know what video Al Qaeda posts. We know photos they post. We know websites they draw on. We’re going to create mathematical values for those things and look for people posting those things,” right? That is a very specific question that technically is quite feasible, but that’s not it.
What the government is proposing is saying, “We’re basically going to decide who’s a good person and bad person and we’re going to look at Twitter to decide the question,” which if you ask a scientist they’re going to say, “No, that’s not what technology is for,” particularly because they’re trying to do this automatically.
Faiza: But all the … You talk about these very nebulous criteria they’re going to put forward in the contract documents and also in the first executive order, but it also asks the system to predict who would commit a terrorist attack, right. Maybe you could talk a little bit … That seems to be a more objective thing that perhaps a system could get its hands on.
Alvaro: Yeah. There was a DoD commissioned study. It was a government commissioned study. I forget who commissioned it. But, it was literally called Rare Events. The core conclusion of the study is that there is no technology available that can accurately predict rare events, I mean terrorism, with any significant degree of confidence and consistency. This is not just me saying this. This is a government funded study reaching that conclusion, number one.
Number two, because I said I think it was from 2009, you can take the word of these scientists and folks in all sorts of wonderful places in industry and the academy saying that’s still true today. This is still too rare to accurately predict.
Yeah, I understand the design. It’s totally understandable that our government wants to keep people out of it that are going to commit terrorist acts, but to try to predict that automatically using machine learning is just not a thing that technology can do in 2017.
Faiza: That seems like a good transition to you, David, since you’re all about rare acts. We’ve talked obviously, and I think we can also recognize that as concerned as we are about the risk of terrorism, that is, fortunately, a very rare event for us here in the United States.
On the other hand, we know that there are these groups out there. We know that they want to do us harm. Oftentimes, they go through and declare their intention to attack the United States or exhort people who follow their ideology to attack the United States. Getting back to the big picture in which we operate, shouldn’t we be looking to mitigate that risk by extreme vetting or bans or things like that?
David: Well, we do mitigate the risk. That’s what the vetting system that we had prior to January 2017 was all about. It was, we are going to give you a review of your history. You have to prove to us who you are and what your intentions are. We do, is what I would argue.
The question is should we go further, and really the question is how much benefit are we going to get from each new investment? It’s another question of whether we should have investment at all. The question is always what’s the bang for the buck, the next buck, not the bang for the first buck. We already spent the first. Now we’re onto the 43 billion and one dollar that we’re spending on this project. How much more are we going to gain by more extreme vetting?
What the numbers show is we have these 13 people who may have radicalized before they came and were sympathizers before they came. How many of those are we going to catch by ramping up the vetting, requiring more information, searching their social media, etc, etc?
Maybe it’s possible that we get one, but the number of people that we would have to deny in order to get to that point would be astronomical. We’d have to deny so many people wrongfully, people who are totally peaceful, have no interest in doing any kind of harm to the United States. That cost is something that we can’t ignore.
So, getting rid of the Diversity Visa Program, which I think there may be arguments for getting rid of it on other grounds, but not terrorism. If you get rid of 50,000 people, immigrants that are coming in, you have to factor in all of the benefits that the United States gains from having these people contributing to the economy, interacting with Americans freely, and so forth.
Really, again, my colleague, he wrote this short post on the executive order and the cost benefit, which is really something that every regulation should go through before it’s released and implemented, is what are the costs from denying these people entry and what are the expected benefits of denying them entry?
He just said, “Let’s just assume that even though none of these people have committed a terrorist attack in the United States, let’s assume that they’re going to commit a terrorist attack at the same rate that every other terrorist that we’ve had, that every other foreign national has committed over the last 43 years. Is it worth it? Is the interest on a base economic level worth the cost?”
It’s not even close. Really, you would need about five times as many terrorist annually than we’ve seen over the last 40 years to make it worth it, and that is just not a level that’s going to be reached anytime soon.
Faiza: Another question I had, you talked a lot about the Muslim ban and this approach of banning certain nationalities. One of the things the administration has cited is the fact that back in 2015, they made, Congress and the previous administration, made changes to the Visa Waiver Program, which basically targeted the same countries. Isn’t there some basis for this idea that there’s a particular threat coming out of those countries that we should be working to mitigate through, whether it’s in this case it was bans, but, then presumably, those very same countries are also going to get extreme vetting. We know they are, in the case of Iraq, for example. Maybe we can talk a little bit about that history.
David: Well, we’re on a continuum here that started a long time ago before 9/11, 9/11 ramped it up, and the Trump Administration is continuing a trend that was preexisting where we do discriminate on national origin in various ways and provide intense vetting of certain groups based on criteria that are vague at best.
But, I would say with respect to that decision, the Obama Administration, really, the only thing that that provision or law requires is that anyone who’s a national in any of these countries get a visa at a consulate. That just requires them to go and apply through the normal channels that are available to anyone else to come to the United States.
The Visa Waiver Program allows you to enter with just a passport. So, if you’re a British national and you’re traveling to the United States, and you’re also a national of Syria, you would have to go and get a visa. That’s discrimination, but it’s not even close to what we’re talking about here, which is a total denial of entry based on solely the fact that you’re a national of a certain country. So, those people would have to be vetted, but they weren’t going to be completely denied entry. And a visa is something that you can use repeatedly in many cases, so those types of visas would allow them to travel back and forth readily. It’s a very different type of proposal that’s come out now with the Trump Administration.
Faiza: In some sense, it’s actually a vote of confidence in the visa vetting procedure, because they were concerned about foreign fighters, so they said, “We’ll go make people who we think might be a risk get a visa.” I think that’s another way to think about it. I just want to thank you both and maybe give you an opportunity to give us some final thoughts. Alvaro, do you-
Alvaro: I feel like I covered a lot of ground. The only thing I want to say is that I think there’s also a lot of stuff going on with respect to leading American businesses that people should pay attention to.
Facebook, for example, decided a couple years ago, actually earlier this year, that it was going to prevent its platform from being used for surveillance. So, it changed its API to prevent law enforcement agencies from automatically just scraping all that information up. If you look at the procurement contracts, they actually call upon the contractor to evade all of those restrictions, by name they specifically talk about Facebook. The more you dig into the digital Muslim ban, the more you dig into this initiative, the more issues there are.
I guess I’ll say if you’re interested in finding out more about this, and if your boss is interested in doing oversight, I would gladly talk to you about this.
Also, I should point out Rachel Levinson-Waldman from the Brennan Center who’s here and is co-leading a coalition with us on this issue. We are looking for folks to conduct oversight.
Faiza: David?
David: I just actually wanted to add a little bit to what Alvaro was talking about. This program of vetting that he’s talking about is really matched with new immigration forms that immigrants are being required to fill out, and those forms ask you to identify whether or not you are a risk to the United States, vague as that might be. If you have a driver’s license, you are a risk to someone in the United States.
These types of vague questions, “Are you going to contribute you to the United States?” What? Whose definition? According to the Trump Administration, grandmothers from any of these six countries were a major threat to the country. So, if you’re going to have a program that’s automatically flagging people, if you answer that question wrongly, the program may come up with, “You lied on your immigration forms,” and if you lie on your immigration forms, that’s a crime. Now, you’re going to jail because of a computer program that said that you lied.
These are the types of … This is the path that we’re going down.
Some of the other things that I think are important, too, to consider about these new immigration forms is that they also ask questions that do have seemingly objective answers, but are actually much more subjective than they first appear.
For example, one of the new questions asks if you were ever detained by a law enforcement official at any point in your entire life. If I was to say, “No, I’ve never been detained by a law enforcement official,” but then I was asked a bunch of questions at the border by someone who has a badge, does that count?
So, the initial inclination to say, “No, I didn’t do this,” or, “I wasn’t ever detained,” is something that could actually end up resulting in you being accused of lying, having your visa denied, having you even prosecuted for lying.
These types of vague questions that are rolled out in a rapid fashion, they were dropped just a few weeks into the Trump Administration, these are the types of things where they just haven’t thought through what the costs are going to be. That’s just another factor to consider when you’re talking about the costs of extreme vetting.