Due to ever-evolving technological advances, manufacturers are connecting consumer goods -- from toys to light bulbs to major appliances -- to the Internet at breakneck speeds. This is the Internet of Things, and it's a security nightmare.
The Internet of Things fuses products with communications technology to make daily life more effortless. Think Amazon's Alexa, which not only answers questions and plays music but allows you to control your home's lights and thermostat. Or the current generation of implanted pacemakers, which can both receive commands and send information to doctors over the Internet.
But like nearly all innovation, there are risks involved. And for products born out of the Internet of Things, this means the risk of having personal information stolen or devices being overtaken and controlled remotely. For devices that affect the world in a direct physical manner -- cars, pacemakers, thermostats -- the risks include loss of life and property.
By developing more advanced security features and building them into these products, hacks can be avoided. The problem is that there is no monetary incentive for companies to invest in the cybersecurity measures needed to keep their products secure. Consumers will buy products without proper security features, unaware that their information is vulnerable. And current liability laws make it hard to hold companies accountable for shoddy software security.
It falls upon lawmakers to create laws that protect consumers. While the US government is largely absent in this area of consumer protection, the state of California has recently steppedin and started regulating the Internet of Things, or "IoT" devices sold in the state -- and the effects will soon be felt worldwide.
California's new SB 327 law, which will take effect in January 2020, requires all "connected devices" to have a "reasonable security feature." The good news is that the term "connected devices" is broadly defined to include just about everything connected to the Internet. The not-so-good news is that "reasonable security" remains defined such that companies trying to avoid compliance can argue that the law is unenforceable.
The legislation requires that security features must be able to protect the device and the information on it from a variety of threats and be appropriate to both the nature of the device and the information it collects. California's attorney general will interpret the law and define the specifics, which will surely be the subject of much lobbying by tech companies.
There's just one specific in the law that's not subject to the attorney general's interpretation: default passwords are not allowed. This is a good thing; they are a terrible security practice. But it's just one of dozens of awful "security" measures commonly found in IoT devices.
This law is not a panacea. But we have to start somewhere, and it is a start.
Though the legislation covers only the state of California, its effects will reach much further. All of us -- in the United States or elsewhere -- are likely to benefit because of the way software is written and sold.
Automobile manufacturers sell their cars worldwide, but they are customized for local markets. The car you buy in the United States is different from the same model sold in Mexico, because the local environmental laws are not the same and manufacturers optimize engines based on where the product will be sold. The economics of building and selling automobiles easily allows for this differentiation.
But software is different. Once California forces minimum security standards on IoT devices, manufacturers will have to rewrite their software to comply. At that point, it won't make sense to have two versions: one for California and another for everywhere else. It's much easier to maintain the single, more secure version and sell it everywhere.
The European General Data Protection Regulation (GDPR), which implemented the annoying warnings and agreements that pop up on websites, is another example of a law that extends well beyond physical borders. You might have noticed an increase in websites that force you to acknowledge you've read and agreed to the website's privacy policies. This is because it is tricky to differentiate between users who are subject to the protections of the GDPR -- people physically in the European Union, and EU citizens wherever they are -- and those who are not. It's easier to extend the protection to everyone.
Once this kind of sorting is possible, companies will, in all likelihood, return to their profitable surveillance capitalism practices on those who are still fair game. Surveillance is still the primary business model of the Internet, and companies want to spy on us and our activities as much as they can so they can sell us more things and monetize what they know about our behavior.
Insecurity is profitable only if you can get away with it worldwide. Once you can't, you might as well make a virtue out of necessity. So everyone will benefit from the California regulation, as they would from similar security regulations enacted in any market around the world large enough to matter, just like everyone will benefit from the portion of GDPR compliance that involves data security.
Most importantly, laws like these spur innovations in cybersecurity. Right now, we have a market failure. Because the courts have traditionally not held software manufacturers liable for vulnerabilities, and because consumers don't have the expertise to differentiate between a secure product and an insecure one, manufacturers have prioritized low prices, getting devices out on the market quickly and additional features over security.
But once a government steps in and imposes more stringent security regulations, companies have an incentive to meet those standards as quickly, cheaply, and effectively as possible. This means more security innovation, because now there's a market for new ideas and new products. We've seen this pattern again and again in safety and security engineering, and we'll see it with the Internet of Things as well.
IoT devices are more dangerous than our traditional computers because they sense the world around us, and affect that world in a direct physical manner. Increasing the cybersecurity of these devices is paramount, and it's heartening to see both individual states and the European Union step in where the US federal government is abdicating responsibility. But we need more, and soon.
Earlier this week, the New York Timesreported that the Russians and the Chinese were eavesdropping on President Donald Trump's personal cell phone and using the information gleaned to better influence his behavior. This should surprise no one. Security experts have beentalkingabout the potential security vulnerabilities in Trump's cell phone use since he became president. And President Barack Obama bristled at -- but acquiesced to -- the security rules prohibiting him from using a "regular" cell phone throughout his presidency.
Three broader questions obviously emerge from the story. Who else is listening in on Trump's cell phone calls? What about the cell phones of other world leaders and senior government officials? And -- most personal of all -- what about my cell phone calls?
There are two basic places to eavesdrop on pretty much any communications system: at the end points and during transmission. This means that a cell phone attacker can either compromise one of the two phones or eavesdrop on the cellular network. Both approaches have their benefits and drawbacks. The NSA seems to prefer bulk eavesdropping on the planet's major communications links and then picking out individuals of interest. In 2016, WikiLeaks published a series of classified documents listing "target selectors": phone numbers the NSA searches for and records. These included senior government officials of Germany -- among them Chancellor Angela Merkel -- France, Japan, andothercountries.
Other countries don't have the same worldwide reach that the NSA has, and must use other methods to intercept cell phone calls. We don't know details of which countries do what, but we know a lot about the vulnerabilities. Insecurities in the phone network itself are so easily exploited that 60 Minutes eavesdropped on a US congressman's phone live on camera in 2016. Back in 2005, unknown attackers targeted the cell phones of many Greek politicians by hacking the country's phone network and turning on an already-installed eavesdropping capability. The NSA even implanted eavesdropping capabilities in networking equipment destined for the Syrian Telephone Company.
Alternatively, an attacker could intercept the radio signals between a cell phone and a tower. Encryption ranges from very weak to possibly strong, depending on which flavor the system uses. Don't think the attacker has to put his eavesdropping antenna on the White House lawn; the Russian Embassy is close enough.
The other way to eavesdrop on a cell phone is by hacking the phone itself. This is the technique favored by countries with less sophisticated intelligence capabilities. In 2017, the public-interest forensics group Citizen Lab uncovered an extensive eavesdropping campaign against Mexican lawyers, journalists, and opposition politicians -- presumably run by the government. Just last month, the same group found eavesdropping capabilities in products from the Israeli cyberweapons manufacturer NSO Group operating in Algeria, Bangladesh, Greece, India, Kazakhstan, Latvia, South Africa -- 45 countries in all.
These attacks generally involve downloading malware onto a smartphone that then records calls, text messages, and other user activities, and forwards them to some central controller. Here, it matters which phone is being targeted. iPhones are harder to hack, which is reflected in the prices companies pay for new exploit capabilities. In 2016, the vulnerability broker Zerodium offered $1.5 million for an unknown iOS exploit and only $200K for a similar Android exploit. Earlier this year, a new Dubai start-up announced even higher prices. These vulnerabilities are resold to governments and cyberweapons manufacturers.
Some of the price difference is due to the ways the two operating systems are designed and used. Apple has much more control over the software on an iPhone than Google does on an Android phone. Also, Android phones are generally designed, built, and sold by third parties, which means they are much less likely to get timely security updates. This is changing. Google now has its own phone -- Pixel -- that gets security updates quickly and regularly, and Google is now trying to pressure Android-phone manufacturers to update their phones more regularly. (President Trump reportedly uses an iPhone.)
Another way to hack a cell phone is to install a backdoor during the design process. This is a real fear; earlier this year, US intelligence officials warned that phones made by the Chinese companies ZTE and Huawei might be compromised by that government, and the Pentagon ordered stores on military bases to stop selling them. This is why China's recommendation that if Trump wanted security, he should use a Huawei phone, was an amusing bit of trolling.
Given the wealth of insecurities and the array of eavesdropping techniques, it's safe to say that lots of countries are spying on the phones of both foreign officials and their own citizens. Many of these techniques are within the capabilities of criminal groups, terrorist organizations, and hackers. If I were guessing, I'd say that the major international powers like China and Russia are using the more passive interception techniques to spy on Trump, and that the smaller countries are too scared of getting caught to try to plant malware on his phone.
It's safe to say that President Trump is not the only one being targeted; so are members of Congress, judges, and other senior officials -- especially because no one is trying to tell any of them to stop using their cell phones (although cell phones still are not allowed on either the House or the Senate floor).
As for the rest of us, it depends on how interesting we are. It's easy to imagine a criminal group eavesdropping on a CEO's phone to gain an advantage in the stock market, or a country doing the same thing for an advantage in a trade negotiation. We've seen governments use these tools against dissidents, reporters, and other political enemies. The Chinese and Russian governments are already targeting the US power grid; it makes sense for them to target the phones of those in charge of that grid.
Unfortunately, there's not much you can do to improve the security of your cell phone. Unlike computer networks, for which you can buy antivirus software, network firewalls, and the like, your phone is largely controlled by others. You're at the mercy of the company that makes your phone, the company that provides your cellular service, and the communications protocols developed when none of this was a problem. If one of those companies doesn't want to bother with security, you're vulnerable.
This is why the current debate about phone privacy, with the FBI on one side wanting the ability to eavesdrop on communications and unlock devices, and users on the other side wanting secure devices, is so important. Yes, there are security benefits to the FBI being able to use this information to help solve crimes, but there are far greater benefits to the phones and networks being so secure that all the potential eavesdroppers -- including the FBI -- can't access them. We can give law enforcement other forensics tools, but we must keep foreign governments, criminal groups, terrorists, and everyone else out of everyone's phones. The president may be taking heat for his love of his insecure phone, but each of us is using just as insecure a phone. And for a surprising number of us, making those phones more private is a matter of national security.
Quantum computing is a new way of computing -- one that could allow humankind to perform computations that are simply impossible using today's computing technologies. It allows for very fast searching, something that would break some of the encryption algorithms we use today. And it allows us to easily factor large numbers, something that would break the RSA cryptosystem for any key length.
This is why cryptographers are hard at work designing and analyzing "quantum-resistant" public-key algorithms. Currently, quantum computing is too nascent for cryptographers to be sure of what is secure and what isn't. But even assuming aliens have developed the technology to its full potential, quantum computing doesn't spell the end of the world for cryptography. Symmetric cryptography is easy to make quantum-resistant, and we're working on quantum-resistant public-key algorithms. If public-key cryptography ends up being a temporary anomaly based on our mathematical knowledge and computational ability, we'll still survive. And if some inconceivable alien technology can break all of cryptography, we still can have secrecy based on information theory -- albeit with significant loss of capability.
At its core, cryptography relies on the mathematical quirk that some things are easier to do than to undo. Just as it's easier to smash a plate than to glue all the pieces back together, it's much easier to multiply two prime numbers together to obtain one large number than it is to factor that large number back into two prime numbers. Asymmetries of this kind -- one-way functions and trap-door one-way functions -- underlie all of cryptography.
To encrypt a message, we combine it with a key to form ciphertext. Without the key, reversing the process is more difficult. Not just a little more difficult, but astronomically more difficult. Modern encryption algorithms are so fast that they can secure your entire hard drive without any noticeable slowdown, but that encryption can't be broken before the heat death of the universe.
With symmetric cryptography -- the kind used to encrypt messages, files, and drives -- that imbalance is exponential, and is amplified as the keys get larger. Adding one bit of key increases the complexity of encryption by less than a percent (I'm hand-waving here) but doubles the cost to break. So a 256-bit key might seem only twice as complex as a 128-bit key, but (with our current knowledge of mathematics) it's 340,282,366,920,938,463,463,374,607,431,768,211,456 times harder to break.
Public-key encryption (used primarily for key exchange) and digital signatures are more complicated. Because they rely on hard mathematical problems like factoring, there are more potential tricks to reverse them. So you'll see key lengths of 2,048 bits for RSA, and 384 bits for algorithms based on elliptic curves. Here again, though, the costs to reverse the algorithms with these key lengths are beyond the current reach of humankind.
This one-wayness is based on our mathematical knowledge. When you hear about a cryptographer "breaking" an algorithm, what happened is that they've found a new trick that makes reversing easier. Cryptographers discover new tricks all the time, which is why we tend to use key lengths that are longer than strictly necessary. This is true for both symmetric and public-key algorithms; we're trying to future-proof them.
Quantum computers promise to upend a lot of this. Because of the way they work, they excel at the sorts of computations necessary to reverse these one-way functions. For symmetric cryptography, this isn't too bad. Grover's algorithm shows that a quantum computer speeds up these attacks to effectively halve the key length. This would mean that a 256-bit key is as strong against a quantum computer as a 128-bit key is against a conventional computer; both are secure for the foreseeable future.
For public-key cryptography, the results are more dire. Shor's algorithm can easily break all of the commonly used public-key algorithms based on both factoring and the discrete logarithm problem. Doubling the key length increases the difficulty to break by a factor of eight. That's not enough of a sustainable edge.
There are a lot of caveats to those two paragraphs, the biggest of which is that quantum computers capable of doing anything like this don't currently exist, and no one knows when -- or even if - we'll be able to build one. We also don't know what sorts of practical difficulties will arise when we try to implement Grover's or Shor's algorithms for anything but toy key sizes. (Error correction on a quantum computer could easily be an unsurmountable problem.) On the other hand, we don't know what other techniques will be discovered once people start working with actual quantum computers. My bet is that we will overcome the engineering challenges, and that there will be many advances and new techniquesbut they're going to take time to discover and invent. Just as it took decades for us to get supercomputers in our pockets, it will take decades to work through all the engineering problems necessary to build large-enough quantum computers.
In the short term, cryptographers are putting considerable effort into designing and analyzing quantum-resistant algorithms, and those are likely to remain secure for decades. This is a necessarily slow process, as both good cryptanalysis transitioning standards take time. Luckily, we have time. Practical quantum computing seems to always remain "ten years in the future," which means no one has any idea.
After that, though, there is always the possibility that those algorithms will fall to aliens with better quantum techniques. I am less worried about symmetric cryptography, where Grover's algorithm is basically an upper limit on quantum improvements, than I am about public-key algorithms based on number theory, which feel more fragile. It's possible that quantum computers will someday break all of them, even those that today are quantum resistant.
If that happens, we will face a world without strong public-key cryptography. That would be a huge blow to security and would break a lot of stuff we currently do, but we could adapt. In the 1980s, Kerberos was an all-symmetric authentication and encryption system. More recently, the GSM cellular standard does both authentication and key distribution -- at scale -- with only symmetric cryptography. Yes, those systems have centralized points of trust and failure, but it's possible to design other systems that use both secret splitting and secret sharing to minimize that risk. (Imagine that a pair of communicants get a piece of their session key from each of five different key servers.) The ubiquity of communications also makes things easier today. We can use out-of-band protocols where, for example, your phone helps you create a key for your computer. We can use in-person registration for added security, maybe at the store where you buy your smartphone or initialize your Internet service. Advances in hardware may also help to secure keys in this world. I'm not trying to design anything here, only to point out that there are many design possibilities. We know that cryptography is all about trust, and we have a lot more techniques to manage trust than we did in the early years of the Internet. Some important properties like forward secrecy will be blunted and far more complex, but as long as symmetric cryptography still works, we'll still have security.
It's a weird future. Maybe the whole idea of number theory-based encryption, which is what our modern public-key systems are, is a temporary detour based on our incomplete model of computing. Now that our model has expanded to include quantum computing, we might end up back to where we were in the late 1970s and early 1980s: symmetric cryptography, code-based cryptography, Merkle hash signatures. That would be both amusing and ironic.
Yes, I know that quantum key distribution is a potential replacement for public-key cryptography. But come on -- does anyone expect a system that requires specialized communications hardware and cables to be useful for anything but niche applications? The future is mobile, always-on, embedded computing devices. Any security for those will necessarily be software only.
There's one more future scenario to consider, one that doesn't require a quantum computer. While there are several mathematical theories that underpin the one-wayness we use in cryptography, proving the validity of those theories is in fact one of the great open problems in computer science. Just as it is possible for a smart cryptographer to find a new trick that makes it easier to break a particular algorithm, we might imagine aliens with sufficient mathematical theory to break all encryption algorithms. To us, today, this is ridiculous. Public- key cryptography is all number theory, and potentially vulnerable to more mathematically inclined aliens. Symmetric cryptography is so much nonlinear muddle, so easy to make more complex, and so easy to increase key length, that this future is unimaginable. Consider an AES variant with a 512-bit block and key size, and 128 rounds. Unless mathematics is fundamentally different than our current understanding, that'll be secure until computers are made of something other than matter and occupy something other than space.
But if the unimaginable happens, that would leave us with cryptography based solely on information theory: one-time pads and their variants. This would be a huge blow to security. One-time pads might be theoretically secure, but in practical terms they are unusable for anything other than specialized niche applications. Today, only crackpots try to build general-use systems based on one-time pads -- and cryptographers laugh at them, because they replace algorithm design problems (easy) with key management and physical security problems (much, much harder). In our alien-ridden science-fiction future, we might have nothing else.
Against these godlike aliens, cryptography will be the only technology we can be sure of. Our nukes might refuse to detonate and our fighter jets might fall out of the sky, but we will still be able to communicate securely using one-time pads. There's an optimism in that.
On May 25, the FBI asked us all to reboot our routers. The story behind this request is one of sophisticated malware and unsophisticated home-network security, and it's a harbinger of the sorts of pervasive threats from nation-states, criminals and hackers that we should expect in coming years.
VPNFilter is a sophisticatedpiece of malware that infects mostly older home and small-office routers made by Linksys, MikroTik, Netgear, QNAP and TP-Link. (For a list of specific models, click here.) It's an impressive piece of work. It can eavesdrop on traffic passing through the router specifically, log-in credentials and SCADA traffic, which is a networking protocol that controls power plants, chemical plants and industrial systems attack other targets on the Internet and destructively "kill" its infected device. It is one of a very few pieces of malware that can survive a reboot, even though that's what the FBI has requested. It has a number of other capabilities, and it can be remotely updated to provide still others. More than 500,000 routers in at least 54 countries have been infected since 2016.
Because of the malware's sophistication, VPNFilter is believed to be the work of a government. The FBI suggested the Russian government was involved for two circumstantial reasons. One, a piece of the code is identical to one found in another piece of malware, called BlackEnergy, that was used in the December 2015 attack against Ukraine's power grid. Russia is believed to be behind that attack. And two, the majority of those 500,000 infections are in Ukraine and controlled by a separate command-and-control server. There might also be classified evidence, as an FBI affidavit in this matter identifies the group behind VPNFilter as Sofacy, also known as APT28 and Fancy Bear. That's the group behind a long list of attacks, including the 2016 hack of the Democratic National Committee.
Two companies, Cisco and Symantec, seem to have been working with the FBI during the past two years to track this malware as it infected ever more routers. The infection mechanism isn't known, but we believe it targets known vulnerabilities in these older routers. Pretty much no one patches their routers, so the vulnerabilities have remained, even if they were fixed in new models from the same manufacturers.
On May 30, the FBI seized control of toknowall.com, a critical VPNFilter command-and-control server. This is called "sinkholing," and serves to disrupt a critical part of this system. When infected routers contact toknowall.com, they will no longer be contacting a server owned by the malware's creators; instead, they'll be contacting a server owned by the FBI. This doesn't entirely neutralize the malware, though. It will stay on the infected routers through reboot, and the underlying vulnerabilities remain, making the routers susceptible to reinfection with a variant controlled by a different server.
If you want to make sure your router is no longer infected, you need to do more than reboot it, the FBI's warning notwithstanding. You need to reset the router to its factory settings. That means you need to reconfigure it for your network, which can be a pain if you're not sophisticated in these matters. If you want to make sure your router cannot be reinfected, you need to update the firmware with any security patches from the manufacturer. This is harder to do and may strain your technical capabilities, though it's ridiculous that routers don't automatically download and install firmware updates on their own. Some of these models probably do not even have security patches available. Honestly, the best thing to do if you have one of the vulnerable models is to throw it away and get a new one. (Your ISP will probably send you a new one free if you claim that it's not working properly. And you should have a new one, because if your current one is on the list, it's at least 10 years old.)
So if it won't clear out the malware, why is the FBI askingus to reboot our routers? It's mostly just to get a sense of how bad the problem is. The FBI now controls toknowall.com. When an infected router gets rebooted, it connects to that server to get fully reinfected, and when it does, the FBI will know. Rebooting will give it a better idea of how many devices out there are infected.
Internet of Things malware isn't new. The 2016 Mirai botnet, for example, created by a lone hacker and not a government, targeted vulnerabilities in Internet-connected digital video recorders and webcams. Other malware has targeted Internet-connected thermostats. Lots of malware targets home routers. These devices are particularly vulnerable because they are often designed by ad hoc teams without a lot of security expertise, stay around in networks far longer than our computers and phones, and have no easy way to patch them.
It wouldn't be surprising if the Russians targeted routers to build a network of infected computers for follow-on cyber operations. I'm sure many governments are doing the same. As long as we allow these insecure devices on the Internet and short of security regulations, there's no way to stop them we're going to be vulnerable to this kind of malware.
And next time, the command-and-control server won't be so easy to disrupt.
Earlier this month, the Pentagon stopped selling phones made by the Chinese companies ZTE and Huawei on military bases because they might be used to spy on their users.
It's a legitimate fear, and perhaps a prudent action. But it's just one instance of the much larger issue of securing our supply chains.
All of our computerized systems are deeply international, and we have no choice but to trust the companies and governments that touch those systems. And while we can ban a few specific products, services or companies, no country can isolate itself from potential foreign interference.
In this specific case, the Pentagon is concerned that the Chinese government demanded that ZTE and Huawei add "backdoors" to their phones that could be surreptitiously turned on by government spies or cause them to fail during some future political conflict. This tampering is possible because the software in these phones is incredibly complex. It's relatively easy for programmers to hide these capabilities, and correspondingly difficult to detect them.
This isn't the first time the United States has taken action against foreign software suspected to contain hidden features that can be used against us. Last December, President Trump signed into law a bill banning software from the Russian company Kaspersky from being used within the US government. In 2012, the focus was on Chinese-made Internet routers. Then, the House Intelligence Committee concluded: "Based on available classified and unclassified information, Huawei and ZTE cannot be trusted to be free of foreign state influence and thus pose a security threat to the United States and to our systems."
Nor is the United States the only country worried about these threats. In 2014, China reportedly banned antivirus products from both Kaspersky and the US company Symantec, based on similar fears. In 2017, the Indian government identified 42 smartphone apps that China subverted. Back in 1997, the Israeli company Check Point was dogged by rumors that its government added backdoors into its products; other of that country's tech companies have been suspected of the same thing. Even al-Qaeda was concerned; ten years ago, a sympathizer released the encryption software Mujahedeen Secrets, claimed to be free of Western influence and backdoors. If a country doesn't trust another country, then it can't trust that country's computer products.
But this trust isn't limited to the country where the company is based. We have to trust the country where the software is written -- and the countries where all the components are manufactured. In 2016, researchers discovered that many different models of cheap Android phones were sending information back to China. The phones might be American-made, but the software was from China. In 2016, researchers demonstrated an even more devious technique, where a backdoor could be added at the computer chip level in the factory that made the chips without the knowledge of, and undetectable by, the engineers who designed the chips in the first place. Pretty much every US technology company manufactures its hardware in countries such as Malaysia, Indonesia, China and Taiwan.
We also have to trust the programmers. Today's large software programs are written by teams of hundreds of programmers scattered around the globe. Backdoors, put there by we-have-no-idea-who, have been discovered in Juniper firewalls and D-Link routers, both of which are US companies. In 2003, someone almost slipped a very clever backdoor into Linux. Think of how many countries' citizens are writing software for Apple or Microsoft or Google.
We can go even farther down the rabbit hole. We have to trust the distribution systems for our hardware and software. Documents disclosed by Edward Snowden showed the National Security Agency installing backdoors into Cisco routers being shipped to the Syrian telephone company. There are fake apps in the Google Play store that eavesdrop on you. Russian hackers subverted the update mechanism of a popular brand of Ukrainian accounting software to spread the NotPetya malware.
I could go on. Supply-chain security is an incredibly complex problem. US-only design and manufacturing isn't an option; the tech world is far too internationally interdependent for that. We can't trust anyone, yet we have no choice but to trust everyone. Our phones, computers, software and cloud systems are touched by citizens of dozens of different countries, any one of whom could subvert them at the demand of their government. And just as Russia is penetrating the US power grid so they have that capability in the event of hostilities, many countries are almost certainly doing the same thing at the consumer level.
We don't know whether the risk of Huawei and ZTE equipment is great enough to warrant the ban. We don't know what classified intelligence the United States has, and what it implies. But we do know that this is just a minor fix for a much larger problem. It's doubtful that this ban will have any real effect. Members of the military, and everyone else, can still buy the phones. They just can't buy them on US military bases. And while the US might block the occasional merger or acquisition, or ban the occasional hardware or software product, we're largely ignoring that larger issue. Solving it borders on somewhere between incredibly expensive and realistically impossible.
Perhaps someday, global norms and international treaties will render this sort of device-level tampering off-limits. But until then, all we can do is hope that this particular arms race doesn't get too far out of control.
Elections serve two purposes. The first, and obvious, purpose is to accurately choose the winner. But the second is equally important: to convince the loser. To the extent that an election system is not transparently and auditably accurate, it fails in that second purpose. Our election systems are failing, and we need to fix them.
Today, we conduct our elections on computers. Our registration lists are in computer databases. We vote on computerized voting machines. And our tabulation and reporting is done on computers. We do this for a lot of good reasons, but a side effect is that elections now have all the insecurities inherent in computers. The only way to reliably protect elections from both malice and accident is to use something that is not hackable or unreliable at scale; the best way to do that is to back up as much of the system as possible with paper.
Recently, there have been two graphic demonstrations of how bad our computerized voting system is. In 2007, the states of California and Ohio conducted audits of their electronic voting machines. Expert review teams found exploitable vulnerabilities in almost every component they examined. The researchers were able to undetectably alter vote tallies, erase audit logs, and load malware on to the systems. Some of their attacks could be implemented by a single individual with no greater access than a normal poll worker; others could be done remotely.
Last year, the Defcon hackers' conference sponsored a Voting Village. Organizers collected 25 pieces of voting equipment, including voting machines and electronic poll books. By the end of the weekend, conference attendees had found ways to compromise every piece of test equipment: to load malicious software, compromise vote tallies and audit logs, or cause equipment to fail.
It's important to understand that these were not well-funded nation-state attackers. These were not even academics who had been studying the problem for weeks. These were bored hackers, with no experience with voting machines, playing around between parties one weekend.
It shouldn't be any surprise that voting equipment, including voting machines, voter registration databases, and vote tabulation systems, are that hackable. They're computers -- often ancient computers running operating systems no longer supported by the manufacturers -- and they don't have any magical security technology that the rest of the industry isn't privy to. If anything, they're less secure than the computers we generally use, because their manufacturers hide any flaws behind the proprietary nature of their equipment.
We're not just worried about altering the vote. Sometimes causing widespread failures, or even just sowing mistrust in the system, is enough. And an election whose results are not trusted or believed is a failed election.
Voting systems have another requirement that makes security even harder to achieve: the requirement for a secret ballot. Because we have to securely separate the election-roll system that determines who can vote from the system that collects and tabulates the votes, we can't use the security systems available to banking and other high-value applications.
We can securely bank online, but can't securely vote online. If we could do away with anonymity -- if everyone could check that their vote was counted correctly -- then it would be easy to secure the vote. But that would lead to other problems. Before the US had the secret ballot, voter coercion and vote-buying were widespread.
We can't, so we need to accept that our voting systems are insecure. We need an election system that is resilient to the threats. And for many parts of the system, that means paper.
Let's start with the voter rolls. We know they've already been targeted. In 2016, someone changed the party affiliation of hundreds of voters before the Republican primary. That's just one possibility. A well-executed attack that deletes, for example, one in five voters at random -- or changes their addresses -- would cause chaos on election day.
Yes, we need to shore up the security of these systems. We need better computer, network, and database security for the various state voter organizations. We also need to better secure the voterregistration websites, with better design and better internet security. We need better security for the companies that build and sell all this equipment.
Multiple, unchangeable backups are essential. A record of every addition, deletion, and change needs to be stored on a separate system, on write-only media like a DVD. Copies of that DVD, or -- even better -- a paper printout of the voter rolls, should be available at every polling place on election day. We need to be ready for anything.
Next, the voting machines themselves. Security researchers agree that the gold standard is a voter-verified paper ballot. The easiest (and cheapest) way to achieve this is through optical-scan voting. Voters mark paper ballots by hand; they are fed into a machine and counted automatically. That paper ballot is saved, and serves as a final true record in a recount in case of problems. Touch-screen machines that print a paper ballot to drop in a ballot box can also work for voters with disabilities, as long as the ballot can be easily read and verified by the voter.
Finally, the tabulation and reporting systems. Here again we need more security in the process, but we must always use those paper ballots as checks on the computers. A manual, post-election, risk-limiting audit varies the number of ballots examined according to the margin of victory. Conducting this audit after every election, before the results are certified, gives us confidence that the election outcome is correct, even if the voting machines and tabulation computers have been tampered with. Additionally, we need better coordination and communications when incidents occur.
It's vital to agree on these procedures and policies before an election. Before the fact, when anyone can win and no one knows whose votes might be changed, it's easy to agree on strong security. But after the vote, someone is the presumptive winner -- and then everything changes. Half of the country wants the result to stand, and half wants it reversed. At that point, it's too late to agree on anything.
The politicians running in the election shouldn't have to argue their challenges in court. Getting elections right is in the interest of all citizens. Many countries have independent election commissions that are charged with conducting elections and ensuring their security. We don't do that in the US.
Instead, we have representatives from each of our two parties in the room, keeping an eye on each other. That provided acceptable security against 20th-century threats, but is totally inadequate to secure our elections in the 21st century. And the belief that the diversity of voting systems in the US provides a measure of security is a dangerous myth, because a few districts can be decisive and there are so few voting-machine vendors.
We candobetter. In 2017, the Department of Homeland Security declared elections to be critical infrastructure, allowing the department to focus on securing them. On 23 March, Congress allocated $380m to states to upgrade election security.
These are good starts, but don't go nearly far enough. The constitution delegates elections to the states but allows Congress to "make or alter such Regulations". In 1845, Congress set a nationwide election day. Today, we need it to set uniform and strict election standards.
Artificial intelligence technologies have the potential to upend the longstanding advantage that attack has over defense on the Internet. This has to do with the relative strengths and weaknesses of people and computers, how those all interplay in Internet security, and where AI technologies might change things.
You can divide Internet security tasks into two sets: what humans do well and what computers do well. Traditionally, computers excel at speed, scale, and scope. They can launch attacks in milliseconds and infect millions of computers. They can scan computer code to look for particular kinds of vulnerabilities, and data packets to identify particular kinds of attacks.
Humans, conversely, excel at thinking and reasoning. They can look at the data and distinguish a real attack from a false alarm, understand the attack as it's happening, and respond to it. They can find new sorts of vulnerabilities in systems. Humans are creative and adaptive, and can understand context.
Computers -- so far, at least -- are bad at what humans do well. They're not creative or adaptive. They don't understand context. They can behave irrationally because of those things.
Humans are slow, and get bored at repetitive tasks. They're terrible at big data analysis. They use cognitive shortcuts, and can only keep a few data points in their head at a time. They can also behave irrationally because of those things.
AI will allow computers to take over Internet security tasks from humans, and then do them faster and at scale. Here are possible AI capabilities:
Discovering new vulnerabilities -- and, more importantly, new types of vulnerabilities in systems, both by the offense to exploit and by the defense to patch, and then automatically exploiting or patching them.
Reacting and adapting to an adversary's actions, again both on the offense and defense sides. This includes reasoning about those actions and what they mean in the context of the attack and the environment.
Abstracting lessons from individual incidents, generalizing them across systems and networks, and applying those lessons to increase attack and defense effectiveness elsewhere.
Identifying strategic and tactical trends from large datasets and using those trends to adapt attack and defense tactics.
That's an incomplete list. I don't think anyone can predict what AI technologies will be capable of. But it's not unreasonable to look at what humans do today and imagine a future where AIs are doing the same things, only at computer speeds, scale, and scope.
Both attack and defense will benefit from AI technologies, but I believe that AI has the capability to tip the scales more toward defense. There will be better offensive and defensive AI techniques. But here's the thing: defense is currently in a worse position than offense precisely because of the human components. Present-day attacks pit the relative advantages of computers and humans against the relative weaknesses of computers and humans. Computers moving into what are traditionally human areas will rebalance that equation.
Roy Amara famously said that we overestimate the short-term effects of new technologies, but underestimate their long-term effects. AI is notoriously hard to predict, so many of the details I speculate about are likely to be wrong -- and AI is likely to introduce new asymmetries that we can't foresee. But AI is the most promising technology I've seen for bringing defense up to par with offense. For Internet security, that will change everything.
Everything online is hackable. This is true for Equifax's data and the federal Office of Personal Management's data, which was hacked in 2015. If information is on a computer connected to the Internet, it is vulnerable.
But just because everything is hackable doesn't mean everything will be hacked. The difference between the two is complex, and filled with defensive technologies, security best practices, consumer awareness, the motivation and skill of the hacker and the desirability of the data. The risks will be different if an attacker is a criminal who just wants credit card details and doesn't care where he gets them from or the Chinese military looking for specific data from a specific place.
The proper question isn't whether it's possible to protect consumer data, but whether a particular site protects our data well enough for the benefits provided by that site. And here, again, there are complications.
In most cases, it's impossible for consumers to make informed decisions about whether their data is protected. We have no idea what sorts of security measures Google uses to protect our highly intimate Web search data or our personal e-mails. We have no idea what sorts of security measures Facebook uses to protect our posts and conversations.
We have a feeling that these big companies do better than smaller ones. But we're also surprised when a lone individual publishes personal data hacked from the infidelity site AshleyMadison.com, or when the North Korean government does the same with personal information in Sony's network.
Think about all the companies collecting personal data about you the websites you visit, your smartphone and its apps, your Internet-connected car -- and how little you know about their security practices. Even worse, credit bureaus and data brokers like Equifax collect your personal information without your knowledge or consent.
So while it might be possible for companies to do a better job of protecting our data, you as a consumer are in no position to demand such protection.
Government policy is the missing ingredient. We need standards and a method for enforcement. We need liabilities and the ability to sue companies that poorly secure our data. The biggest reason companies don't protect our data online is that it's cheaper not to. Government policy is how we change that.
This essay appeared as half of a point/counterpoint with Priscilla Regan, in a CQ Researcher report titled "Privacy and the Internet."
For over a decade, civil libertarians have been fighting government mass surveillance of innocent Americans over the Internet. We've just lost an important battle. On January 18, President Trumpsigned the renewal of Section 702, domestic mass surveillance became effectively a permanent part of US law.
Section 702 was initially passed in 2008, as an amendment to the Foreign Intelligence Surveillance Act of 1978. As the title of that law says, it was billed as a way for the NSA to spy on non-Americans located outside the United States. It was supposed to be an efficiency and cost-saving measure: the NSA was already permitted to tap communications cables located outside the country, and it was already permitted to tap communications cables from one foreign country to another that passed through the United States. Section 702 allowed it to tap those cables from inside the United States, where it was easier. It also allowed the NSA to request surveillance data directly from Internet companies under a program called PRISM.
The problem is that this authority also gave the NSA the ability to collect foreign communications and data in a way that inherently and intentionally also swept up Americans' communications as well, without a warrant. Other law enforcement agencies are allowed to ask the NSA to search those communications, give their contents to the FBI and other agencies and then lie about their origins in court.
In 1978, after Watergate had revealed the Nixon administration's abuses of power, we erected a wall between intelligence and law enforcement that prevented precisely this kind of sharing of surveillance data under any authority less restrictive than the Fourth Amendment. Weakening that wall is incredibly dangerous, and the NSA should never have been given this authority in the first place.
Arguably, it never was. The NSA had been doing this type of surveillance illegally for years, something that was first made public in 2006. Section 702 was secretly used as a way to paper over that illegal collection, but nothing in the text of the later amendment gives the NSA this authority. We didn't know that the NSA was using this law as the statutory basis for this surveillance until Edward Snowden showed us in 2013.
Civil libertarians have been battling this law in both Congress and the courts ever since it was proposed, and the NSA's domestic surveillance activities even longer. What this most recent vote tells me is that we've lost that fight.
Section 702 was passed under George W. Bush in 2008, reauthorized under Barack Obama in 2012, and now reauthorized again under Trump. In all three cases, congressional support was bipartisan. It has survived multiple lawsuits by the Electronic Frontier Foundation, the ACLU, and others. It has survived the revelations by Snowden that it was being used far more extensively than Congress or the public believed, and numerous public reports of violations of the law. It has even survived Trump's belief that he was being personally spied on by the intelligence community, as well as any congressional fears that Trump could abuse the authority in the coming years. And though this extension lasts only six years, it's inconceivable to me that it will ever be repealed at this point.
So what do we do? If we can't fight this particular statutory authority, where's the new front on surveillance? There are, it turns out, reasonable modifications that target surveillance more generally, and not in terms of any particular statutory authority. We need to look at US surveillance law more generally.
First, we need to strengthen the minimization procedures to limit incidental collection. Since the Internet was developed, all the world's communications travel around in a single global network. It's impossible to collect only foreign communications, because they're invariably mixed in with domestic communications. This is called "incidental" collection, but that's a misleading name. It's collected knowingly, and searched regularly. The intelligence community needs much stronger restrictions on which American communications channels it can access without a court order, and rules that require they delete the data if they inadvertently collect it. More importantly, "collection" is defined as the point the NSA takes a copy of the communications, and not later when they search their databases.
Second, we need to limit how other law enforcement agencies can use incidentally collected information. Today, those agencies can query a database of incidental collection on Americans. The NSA can legally pass information to those other agencies. This has to stop. Data collected by the NSA under its foreign surveillance authority should not be used as a vehicle for domestic surveillance.
The most recent reauthorization modified this lightly, forcing the FBI to obtain a court order when querying the 702 data for a criminal investigation. There are still exceptions and loopholes, though.
Third, we need to end what's called "parallel construction." Today, when a law enforcement agency uses evidence found in this NSA database to arrest someone, it doesn't have to disclose that fact in court. It can reconstruct the evidence in some other manner once it knows about it, and then pretend it learned of it that way. This right to lie to the judge and the defense is corrosive to liberty, and it must end.
Pressure to reform the NSA will probably first come from Europe. Already, European Union courts have pointed to warrantless NSA surveillance as a reason to keep Europeans' data out of US hands. Right now, there is a fragile agreement between the EU and the United States -- called "Privacy Shield" -- that requires Americans to maintain certain safeguards for international data flows. NSA surveillance goes against that, and it's only a matter of time before EU courts start ruling this way. That'll have significant effects on both government and corporate surveillance of Europeans and, by extension, the entire world.
Further pressure will come from the increased surveillance coming from the Internet of Things. When your home, car, and body are awash in sensors, privacy from both governments and corporations will become increasingly important. Sooner or later, society will reach a tipping point where it's all too much. When that happens, we're going to see significant pushback against surveillance of all kinds. That's when we'll get new laws that revise all government authorities in this area: a clean sweep for a new world, one with new norms and new fears.
It's possible that a federal court will rule on Section 702. Although there have been many lawsuits challenging the legality of what the NSA is doing and the constitutionality of the 702 program, no court has ever ruled on those questions. The Bush and Obama administrations successfully argued that defendants don't have legal standing to sue. That is, they have no right to sue because they don't know they're being targeted. If any of the lawsuits can get past that, things might change dramatically.
Meanwhile, much of this is the responsibility of the tech sector. This problem exists primarily because Internet companies collect and retain so much personal data and allow it to be sent across the network with minimal security. Since the government has abdicated its responsibility to protect our privacy and security, these companies need to step up: Minimize data collection. Don't save data longer than absolutely necessary. Encrypt what has to be saved. Well-designed Internet services will safeguard users, regardless of government surveillance authority.
For the rest of us concerned about this, it's important not to give up hope. Everything we do to keep the issue in the public eye -- and not just when the authority comes up for reauthorization again in 2024 -- hastens the day when we will reaffirm our rights to privacy in the digital age.
In August, four US Senators introduced a bill designed to improve Internet of Things (IoT) security. The IoT Cybersecurity Improvement Act of 2017 is a modest piece of legislation. It doesn't regulate the IoT market. It doesn't single out any industries for particular attention, or force any companies to do anything. It doesn't even modify the liability laws for embedded software. Companies can continue to sell IoT devices with whatever lousy security they want.
What the bill does do is leverage the government's buying power to nudge the market: any IoT product that the government buys must meet minimum security standards. It requires vendors to ensure that devices can not only be patched, but are patched in an authenticated and timely manner; don't have unchangeable default passwords; and are free from known vulnerabilities. It's about as low a security bar as you can set, and that it will considerably improve security speaks volumes about the current state of IoT security. (Full disclosure: I helped draft some of the bill's security requirements.)
The bill would also modify the Computer Fraud and Abuse and the Digital Millennium Copyright Acts to allow security researchers to study the security of IoT devices purchased by the government. It's a far narrower exemption than our industry needs. But it's a good first step, which is probably the best thing you can say about this legislation.
However, it's unlikely this first step will even be taken. I am writing this column in August, and have no doubt that the bill will have gone nowhere by the time you read it in October or later. If hearings are held, they won't matter. The bill won't have been voted on by any committee, and it won't be on any legislative calendar. The odds of this bill becoming law are zero. And that's not just because of current politics -- I'd be equally pessimistic under the Obama administration.
But the situation is critical. The Internet is dangerous -- and the IoT gives it not just eyes and ears, but also hands and feet. Security vulnerabilities, exploits, and attacks that once affected only bits and bytes now affect flesh and blood.
Markets, as we've repeatedly learned over the past century, are terrible mechanisms for improving the safety of products and services. It was true for automobile, food, restaurant, airplane, fire, and financial-instrument safety. The reasons are complicated, but basically, sellers don't compete on safety features because buyers can't efficiently differentiate products based on safety considerations. The race-to-the-bottom mechanism that markets use to minimize prices also minimizes quality. Without government intervention, the IoT remains dangerously insecure.
The US government has no appetite for intervention, so we won't see serious safety and security regulations, a new federal agency, or better liability laws. We might have a better chance in the EU. Depending on how the General Data Protection Regulation on data privacy pans out, the EU might pass a similar security law in 5 years. No other country has a large enough market share to make a difference.
Sometimes we can opt out of the IoT, but that option is becoming increasingly rare. Last year, I tried and failed to purchase a new car without an Internet connection. In a few years, it's going to be nearly impossible to not be multiply connected to the IoT. And our biggest IoT security risks will stem not from devices we have a market relationship with, but from everyone else's cars, cameras, routers, drones, and so on.
We can try to shop our ideals and demand more security, but companies don't compete on IoT safety -- and we security experts aren't a large enough market force to make a difference.
We need a Plan B, although I'm not sure what that is. Comment if you have any ideas.
This essay previously appeared in the September/October issue of IEEE Security & Privacy.