February 22, 2017

Engineering around social media border searches

The latest news is that the U.S. Department of Homeland Security is considering a requirement, while passing through a border checkpoint, to inspect a prospective visitor’s “online presence”. That means immigration officials would require users to divulge their passwords to Facebook and other such services, which the agent might then inspect, right there, at the border crossing. This raises a variety of concerns, from its chilling impact on freedom of speech to its being an unreasonable search or seizure, nevermind whether an airport border agent has the necessary training to make such judgments, much less the time to do it while hundreds of people are waiting in line to get through.

Rather than conduct a serious legal analysis, however, I want to talk about technical countermeasures. What might Facebook or other such services do to help defend their users as they pass a border crossing?

Fake accounts. It’s certainly feasible today to create multiple accounts for yourself, giving up the password to a fake account rather than your real account. Most users would find this unnecessarily cumbersome, and the last thing Facebook or anybody else wants is to have a bunch of fake accounts running around. It’s already a concern when somebody tries to borrow a real person’s identity to create a fake account and “friend” their actual friends.

Duress passwords. Years ago, my home alarm system had the option to have two separate PINs. One of them would disable the alarm as normal. The other would sound a silent alarm, summoning the police immediately while making it seem like I disabled the alarm. Let’s say Facebook supported something similar. You enter the duress password, then Facebook locks out your account or switches to your fake account, as above.

Temporary lockouts. If you know you’re about to go through a border crossing, you could give a duress password, as above, or you could arrange an account lockout in advance. You might, for example, designate ten trusted friends, where any five must declare that the lockout is over. Absent those declarations, your account would remain locked, and there would be no means for you to be coerced into giving access to your own account.

Temporary sanitization. Absent any action from Facebook, the best advice today for somebody about to go through a border crossing is to sanitize their account before going through. That means attempting to second-guess what border agents are looking for and delete it in advance. Facebook might assist this by providing search features to allow users to temporarily drop friends, temporarily delete comments or posts with keywords in them, etc. As with the temporary lockouts, temporary sanitization would need to have a restoration process that could be delegated to trusted friends. Once you give the all-clear, everything comes back again.

User defense in bulk. Every time a user, going through a border crossing, exercises a duress password, that’s an unambiguous signal to Facebook. Even absent such signals, Facebook would observe highly unusual login behavior coming from those specific browsers and IP addresses. Facebook could simply deny access to its services from government IP address blocks. While it’s entirely possible for the government to circumvent this, whether using Tor or whatever else, there’s no reason that Facebook needs to be complicit in the process.

So is there a reasonable alternative?

While it’s technically feasible for the government to require that Facebook give it full “backdoor” access to each and every account so it can render threat judgments in advance, this would constitute the most unreasonable search and seizure in the history of that phrase. Furthermore, if and when it became common knowledge that such unreasonable seizures were commonplace, that would be the end of the company. Facebook users have an expectation of privacy and will switch to other services if Facebook cannot protect them.

Wouldn’t it be nice if there was some less invasive way to support the government’s desire for “extreme vetting”? Can we protect ordinary users’ privacy while still enabling the government to intercept people who intend harm to our country? We certainly must assume that an actual bona fide terrorist is going to have no trouble creating a completely clean online persona to use while crossing a border. They can invent wholesome friends with healthy children sharing silly videos of cute kittens. While we don’t know too much about our existing vetting strategies to distinguish tourists from terrorists, we have to assume that the process involves the accumulation of signals and human intelligence, and other painstaking efforts by professional investigators to protect our country from harm. It’s entirely possible that they’re already doing a good job.

A response to the National Association of Secretaries of State

NASS logo
Election administration in the United States is largely managed state-by-state, with a small amount of Federal involvement. This generally means that each state’s chief election official is that state’s Secretary of State. Their umbrella organization, the National Association of Secretaries of State, consequently has a lot of involvement in voting issues, and recently issued a press release concerning voting system security that was remarkably erroneous. What follows is a point-by-point commentary on their press release.

To date, there has been no indication from national security agencies to states that any specific or credible threat exists when it comes to cyber security and the November 2016 general election.

Unfortunately, we now know that it appears that Russia broke into the DNC’s computers and leaked emails with clear intent to influence the U.S. presidential election (see, e.g., the New York Times’s article on July 26: “Why Security Experts Think Russia was Behind the DNC Breach”). It’s entirely reasonable to extrapolate from this that they may be willing to conduct further operations with the same goals, meaning that it’s necessary to take appropriate steps to mitigate against such attacks, regardless of the level of specificity of available intel.

However, as a routine part of any election cycle, Secretaries of State and their local government counterparts work with federal partners, such as the U.S. Election Assistance Commission (EAC) and the National Institute of Standards and Technology (NIST), to maintain rigorous testing and certification standards for voting systems. Risk management practices and controls, including the physical handling and storage of voting equipment, are important elements of this work.

Expert analyses of current election systems (largely conducted ten years ago in California, Ohio, and Florida) found a wide variety of security problems. While some states have responded to these issues by replacing the worst paperless electronic voting systems, other states, including several “battleground” states, continue to use unacceptably insecure systems.

State election offices also proactively utilize election IT professionals and security experts to regularly review, identify and address any vulnerabilities with systems, including voter registration databases and election night reporting systems (which display the unofficial tallies that are ultimately verified via statewide canvassing).

The implication here is that all state election officials have addressed known vulnerabilities. This is incorrect. While some states have been quite proactive, other states have done nothing of the sort.

A national hacking of the election is highly improbable due to our unique, decentralized process.

Security vulnerabilities have nothing to do with probabilities. They instead have to do with a cost/benefit analysis on the part of the attacker. An adversary doesn’t have to attack all 50 states. All they have to do is tamper with the “battleground” states where small shifts in the vote can change the outcome for the whole state.

Each state and locality conducts its own system of voting, complete with standards and security requirements for equipment and software. Most states publicly conduct logic and accuracy testing of their machines prior to the election to ensure that they are working and tabulating properly, then they are sealed until Election Day to prevent tampering.

So-called “logic and accuracy testing” varies from location to location, but most boil down to casting a small number of votes for each candidate, on a handful of machines, and making sure they’re all there in a mock tally. Similarly, local election officials will have procedures in place to make sure machines are properly “zeroed”. Computer scientists refer to these as “sanity tests”, in that if the system fails, then something is obviously broken. If these tests pass, they say nothing about the sort of tampering that a sophisticated nation-state adversary might conduct.

Some election officials conduct more sophisticated “parallel testing”, where some voting equipment is pulled out of general service and is instead set up in a mock precinct, on election day, where mock voters cast seemingly real ballots. These machines would have a harder time distinguishing whether they were in “test” versus “production” conditions. But what happens if the machines fail the parallel test? By then, the election is over, the voters are gone, and there’s potentially no way to reconstruct the intent of the voters.

Furthermore, electronic voting machines are not Internet-based and do not connect to each other online.

This is partially true. Electronic voting systems do connect to one another through in-precinct local networks or through the motion of memory cards of various sorts. They similarly connect to election management systems before the start of the election (when they’re loaded with ballot definitions) and after the end of the election (for backups, recounts, inventory control, and/or being cleared prior to subsequent elections). All of these “touch points” represent opportunities for malware to cross the “air gap” boundaries. We built attacks like these a decade ago as part of the California Top to Bottom Review, showing how malware could spread “virally” to an entire county’s fleet of voting equipment. Attacks like these require a non-trivial up-front engineering effort, plus additional effort for deployment, but these efforts are well within the capabilities of a nation-state adversary.

Following the election, state and local jurisdictions conduct a canvass to review vote counting, ultimately producing the election results that are officially certified. Post-election audits help to further guard against deliberate manipulation of the election, as well as unintentional software, hardware or programming problems.

Post-election audits aren’t conducted at all in some jurisdictions, and would likely be meaningless against the sort of adversary we’re talking about. If a paperless electronic voting system was hacked, there might well be forensic evidence that the attackers left behind, but such evidence would be a challenge to identify quickly, particularly in the charged atmosphere of a disputed election result.

We look forward to continued information-sharing with federal partners in order to evaluate cyber risks, and respond to them accordingly, as part of ongoing state election emergency preparedness planning for November.

“Emergency preparedness” is definitely the proper way to consider the problem. Just as we must have contingency plans for all sorts of natural phenomena, like hurricanes, we must also be prepared for man-made phenomena, where we might be unable to reconstruct an election tally that accurately represents the will of the people.

The correct time to make such plans is right now, before the election. Since it’s far too late to decommission and replace our insecure equipment, we must instead plan for rapid responses, such as quickly printing single-issue paper ballots, bringing voters back to the polls, and doing it all over again. If such plans are made now, their very existence changes the cost/benefit equation for our adversaries, and will hopefully dissuade these adversaries from acting.

Election security as a national security issue

We recently learned that Russian state actors may have been responsible for the DNC emails recently leaked to Wikileaks. Earlier this spring, once they became aware of the hack, the DNC hired Crowdstrike, an incident response firm. The New York Times reports:

Preliminary conclusions were discussed last week at a weekly cyberintelligence meeting for senior officials. The Crowdstrike report, supported by several other firms that have examined the same bits of code and telltale “metadata” left on documents that were released before WikiLeaks’ publication of the larger trove, concludes that the Federal Security Service, known as the F.S.B., entered the committee’s networks last summer.

President Obama added that “on a regular basis, [the Russians] try to influence elections in Europe.” For the sake of this blog piece, and it’s not really a stretch, let’s take it as a given that foreign nation-state actors including Russia have a large interest in the outcome of U.S. elections and are willing to take all sorts of unseemly steps to influence what happens here. Let’s take it as a given that this is undesirable and talk about how we might stop it.

It’s bad enough to see foreign actors leaking emails with partisan intent. To make matters worse,  Bruce Schneier in a Washington Post op-ed and many other security experts in the past have been worried about our voting systems themselves being hacked. How bad could this get? Several companies are now offering Internet-based voting systems alongside apparently unfounded claims as to their security. In one example, Washington D.C. looked at using one such system for its local elections and had a “pilot” in 2010, wherein the University of Michigan’s Alex Halderman and his students found and exploited significant security vulnerabilities. Had this system been used in a real election, any foreign nation-state actor could have done the same. Luckily, these systems aren’t widely used.

How vulnerable are our nation’s election systems, as they’ll be used this November 2016, to being manipulated by foreign nation-state actors? The answer depends on how close the election will be. Consider Bush v. Gore in 2000. If an attacker, knowing it would be a very close election, had found a way to specifically manipulate the outcome in Florida, then their attack could well have had a decisive impact. Of course, predicting election outcomes is as much an art as a science, so an attacker would need to hedge their bets and go after the voting systems in multiple “battleground” states. Conversely, there’s no point in going after highly polarized states, where small changes will have no decisive impact. As an attacker, you want to leave a minimal footprint.

How good are we at defending ourselves? Will cyber attacks on current voting systems leave evidence that can be detected prior to our elections? Let’s consider the possible attacks and how our defenses might respond.

Voter de-registration: The purpose of a many attacks is simply to break things. Applied with partisan intent, you’d want to break things for one party more than the other. The easiest attack would be to hack a voter registration system, deleting voters who you believe are likely to support the candidate you don’t like. For voters who have registered for a political party, you know everything you need to know for who to delete. For independent voters you can probabilistically infer a their political opinions based on how their local precinct votes and on other demographic variables. (Political scientists do this sort of thing all the time.) Selectively destroying voter registration databases is likely to be recoverable. Such voters could demand to vote “provisional ballots” and those ballots would get counted as normal, once the voter registration databases were restored.

Vote flipping: A nastier attack would require an attacker to access the computers inside DRE voting systems. (“Direct recording electronic” systems are typically touch-screen computers with no voter-verifiable paper trail. The only record of a voter’s ballot is stored electronically, inside the computer.) These voting systems are typically not connected to the Internet, although they do connect to election management computers, and those sometimes use modems to gather data from remote precincts. (Details vary from state to state and even county to county.) From the perspective of a nation-state cyber attacker, a modem might as well be a direct connection to the Internet. Once you can get malware into one of these election management computers, you can delete or flip votes. If you’re especially clever, you can use the occasional connections from these election management computers to the voting machines and corrupt the voting machines themselves. (We showed how to do these sort of viral attacks as part of the California Top to Bottom Review in 2007.)

With paperless DRE systems, attacked by a competent nation-state actor, there will be no reason to believe any of the electronic records are intact, and a competent attacker would presumably also be good enough to clean up on their way out, so there wouldn’t necessarily even be any evidence of the attack.

The good news is that paperless DRE systems are losing market share and being replaced slowly-but-surely with several varieties of paper-ballot systems (some hand-marked and electronically scanned, others machine-marked). A foreign nation-state adversary can’t reach across the Internet and change what’s printed on a piece of paper, which means that a post-election auditing strategy to compare the electronic results to the paper results can efficiently detect (and thus deter) electronic tampering.

Where would an adversary attack? The most bang-for-the-buck for a foreign nation-state bent on corrupting our election would be to find a way to tamper with paperless DRE voting systems in a battleground state. So where then? Check out the NYT’s interactive “paths to the White House” page, wherein you can play “what-if” games on which states might have what impact in the Electoral College. The top battleground state is Florida, but thanks in part to the disastrous 2006 election in Florida’s 13th Congressional district, Florida dumped its DRE voting systems for optically scanned paper ballots; it would be much harder for an adversarial cyber attack to go undetected. What about other battleground states? Following the data in the Verified Voting website, Pennsylvania continues to use paperless DREs as does Georgia. Much of Ohio uses DRE systems with “toilet paper roll” printers, where voters are largely unable to detect if anything is printed incorrectly, so we’ll lump them in with the paperless states. North Carolina uses a mix of technologies, some of which are more vulnerable than others. So let’s say the Russians want to rig the election for Trump. If they could guarantee a Trump win in Pennsylvania, Georgia, Ohio, and North Carolina, then a Florida victory could put Trump over the top. Even without conspiracy theories, Florida will still be an intensely fought battleground state, but we don’t need a foreign government making it any worse.

So what should these sensitive states do in the short term? At this point, it’s far too late to require non-trivial changes in election technologies or even most procedures. They’re committed to what they’ve got and how they’ll use it. We could imagine requiring some essential improvements (security patches and updates installed, intrusion detection and monitoring equipment installed, etc.) and even some sophisticated analyses (e.g., pulling voting machines off the line and conducting detailed / destructive analyses of their internal state, going beyond the weak tamper-protection mechanisms presently in place). Despite all of this, we could well end up in a scenario where we conclude that we have unreliable or tampered election data and cannot use it to produce a meaningful vote tally.

Consider also that all an adversary needs to do is raise enough doubt that the loser has seemingly legitimate grounds to dispute the result. Trump is already suggesting that this November’s election might be rigged, without any particular evidence to support this conjecture. This makes it all the more essential that we have procedures that all parties can agree to for recounts, for audits, and for what to do when those indicate discrepancies.

In case of emergency, break glass. If we’re facing a situation where we see tampering on a massive scale, we could end up in a crisis far worse than Florida after the Bush/Gore election of 2000. If we do nothing until after we find problems, every proposed solution will be tinted with its partisan impact, making it difficult to reach any sort of procedural consensus. Nobody wants to imagine a case where our electronic voting systems have been utterly compromised, but if we establish processes and procedures, in advance, for dealing with these contingencies, such as commissioning paper ballots and rerunning the elections in impacted areas, we will disincentivize foreign election adversaries and preserve the integrity of our democracy.

(Addendum: contingency planning was exactly the topic of discussion after Hurricane Sandy disrupted elections across the Northeast in November 2012. It would be useful to revisit whatever changes were made then, in light of the new threat landscape we have today.)

Related reading: