“[E]ncryption technologies are the most important technological breakthrough in the last one thousand years.”
--Professor Lawrence Lessig, Code and Other Laws of Cyberspace 36 (1999).
Needless to say, the Internet and other computer and communications technologies are radically altering the ways in which we communicate and exchange information. However, along with the speed, efficiency, and cost-saving benefits of the “digital revolution” come new challenges to the security and privacy of communications and information traversing the global communications infrastructure.
In response to these challenges, the security mechanisms of traditional paper-based communications media – envelopes and locked filing cabinets – are being replaced by cryptographic security techniques. Through the use of a technique called encryption, or cryptography, communication and information stored and transmitted by computers can be protected against interception to a very high degree. Until recently, there was little non-governmental demand for encryption capabilities. Modern encryption technology – a mathematical process involving the use of formulas (or algorithms) – was traditionally deployed most widely to protect the confidentiality of military and diplomatic communications. However, with the advent of the computer revolution, and recent innovations in the science of encryption, a new market for cryptographic products has developed. Electronic communications are now widely used in the civilian sector and have become an integral component of the global economy. Computers store and exchange an ever-increasing amount of highly personal information, including medical and financial data. In this electronic environment, the need for privacy-enhancing technologies is apparent.
II. How Does Encryption Work?
The object of this section is to provide a basic overview of how modern encryption technology works. To understand precisely the mechanics of encryption would require months and a solid foundation in advanced mathematics. Given that this section is being written only to provide a background for understanding the legal issues surrounding encryption, we will only skim over the main features of encryption.
Before we jump into encryption, let’s start with the concept of data interception. Data interception occurs when an unauthorized person is able to seize a piece of information that is being sent from person A to person B before B receives that information. Let’s consider an example. Suppose that I wanted to send Bill a message that reads, “HELLO.” The problem is that a hacker might intercept my message before Bill receives the message and be able to read what I am trying to communicate. Now, in a message as simple as “HELLO,” this interception may seem rather harmless. But suppose that instead of just sending Bill a “HELLO” message, I was sending Bill my credit card number for an online purchase of books. When we think about the possibility of credit card numbers getting into the wrong hands, we see the potential seriousness of data interception.
How can I protect my messages from being intercepted by unauthorized people? Unfortunately, because information on the Internet passes through public wires, it is not possible to ensure that unauthorized users will not be able to intercept data. However, even if these unauthorized users are able to intercept data, it is possible to mask the data so that hackers and other snoopers cannot read the data even if they intercept it. This process of translating data into a code that makes it more difficult for unauthorized users to read is called encryption, or cryptography. For example, instead of sending Bill the message, “HELLO,” I could have sent him the message, “8-5-12-12-15.” Now, the message “8-5-12-12-15” seems like gibberish and may not be understood by a hacker, even if she is able to intercept my message. But suppose that along with the “8-5-12-12-15” message, I let Bill know that each number in my message should be replaced by the corresponding letter in the alphabet. Given this information, Bill will be able to decrypt, or decipher, my message.
The information that Bill used to decipher my encrypted message is called a key. Keep in mind here that I used the same key to encrypt the message (i.e., numbers in the message to be replaced by the corresponding letter in the alphabet) as Bill used to decrypt the message. This is called the single-key system of encryption. Until the mid-1970s, encryption operated under this system – i.e., the same key that was used to encrypt a message was also used to decrypt it. However, there was a great risk inherent in this single-key system: if you lose that single key, then you risk losing all the information that was encrypted by that key. For example, if the key that “numbers should be replaced by the corresponding letters of the alphabet” gets lost, and it is impossible to decipher the code without the key, then Bill will never be able to decrypt the message that I sent to him. Furthermore, the single-key system always required that you transport the key needed to decrypt the encrypted message, which added a further risk that the key would be lost during transport.
The pitfalls of the single-key system led computer scientists searching for an alternative. And in the mid-1970s, Whitfield Diffie and Martin Hellman devised a system that is still used today by most encryption programs. Their system, the dual-key system, relied on two keys – a public key and a private key. Under the dual-key system, what is encrypted by the public key can only be decrypted by the corresponding private key. Furthermore, the system is set up such that even with one key, it is impossible to infer or derive the other. So how does this system work? Suppose that I wanted to send a message to Bill again. Under the dual-key system, Bill would have a key that he can make publicly available (the public key). Therefore, I can take my message and encrypt it using Bill’s public key. I then send the encrypted message to Bill. The message that I have sent (which has been encrypted by Bill’s public key), can only be decrypted by using Bill’s private key. And since Bill is the only one who has his private key, no one other than Bill (and anybody else who he authorizes to use his private key) can read the encrypted message that I have sent to him. Once Bill receives my encrypted message, he can decrypt it with his private key and read the message.
Under the process just described, it is clear that only Bill will be able to read the message that I have sent to him. But how will Bill know that I was the one who sent the message? After all, anyone can encrypt a message using Bill’s public key. One way around this problem is before I send the message that I have encrypted with Bill’s public key, I can encrypt it first with my private key. Then, when Bill receives my encrypted message, he can first decrypt it with my public key, and then decrypt it again with his private key. After the first decryption, Bill can be sure that I was the one who sent him the message. And after the second decryption, Bill can be sure that only he can read the content of the message.
The dual-key system’s advantage over the single-key system is clear. Under the dual-key system, Bill and I did not need to send a single key back and forth. I simply used Bill’s public key to encrypt the message, and Bill used his corresponding private key to decrypt it. There was no risk that the key used to encrypt the message would get lost. Hence, the risk of lost information under the dual-key system is much lower than under the single-key system.
In order to get a public and private key to encrypt messages, you need to have a software program. Currently, the most popular dual-key software programs are called RSA and PGP (Pretty Good Privacy).
III. A Survey of Issues Surrounding Encryption
Modern encryption, as we have just discussed, is achieved with algorithms that use keys to encrypt and decrypt messages by turning text or other data into digital code and then by restoring it to its original form. The length of the key determines the code’s security level – the longer the key, the more secure the code. To decipher an encrypted message without access to a key, a person would need to try every possible key. Computer keys are made of bits of information, binary units of information that can have the value of 0 or 1. Therefore, an 8-bit long key has 256 (28) possible values. A 56-bit key creates 72 quadrillion (72,000,000,000,000,000) possible combinations. Without the key to crack a 56-bit encrypted message, a person would have to resort to the so-called brute-force method to decrypt the message – i.e., try out every single one of the 72 quadrillion possible combinations. If the key is 128 bits long, attempting to crack the code without the key would be 4.7 sextillion (4,700,000,000,000,000,000,000) times more difficult than cracking a 56-bit key (which itself has 72 quadrillion possible combinations)! Given the current power of computers, experts consider that a 56-bit key could be cracked by using the brute-force method in 10 million hours of computer time (14,000 computers used around the clock for 4 months). However, a 128-bit key is not considered crackable. Until 1996, the U.S. government considered anything stronger than a 40-bit encryption to be a “munition” – hence, the export of any piece of information with that level of encryption was illegal. Since then, the government has relaxed its standards and allows the export of 56-bit encryption, with some restrictions. 128-bit encryption has now emerged as the standard of illegality.
The government’s restrictions on 128-bit encryption, however, has led to opposition from many corners, including software companies that are worried that restrictions will impede the growth of Internet commerce and organizations, such as the ACLU, that are concerned that the restrictions will intrude on individuals’ privacy rights. In response, the government, led by the Department of Justice (DOJ) and Federal Bureau of Investigations (FBI) argue that unbreakable encryption will destroy law enforcement officials’ ability to fight crime and prevent terrorism, and will ultimately endanger national security. How to balance these competing interests is an issue that is still perplexing policymakers and legislators. In the following sections, we will lay down the main points that these competing groups have put forth.
IV. Arguments For Restrictions on Cryptography – Law Enforcement
The government’s fear of unbreakable encryption is simple – if encryption is used by the wrong people, then law enforcement officials will be powerless to stop those people from committing crimes of extraordinary degree. Law enforcement officials rely on electronic surveillance as a tool of utmost importance in terrorism cases and in many criminal investigations, especially those involving serious and violent crime, terrorism, espionage, organized crime, drug-trafficking, corruption and fraud. Unbreakable encryption will render such methods of electronic surveillance useless; even if law enforcement officials intercept the communication made between criminals, they will be unable to decipher the encrypted messages. As a result, drug lords, spies, and even violent gangs will be able to communicate about their crimes and their conspiracies with impunity.
According to FBI Director Louis Freeh, this is not a problem that will begin sometime in the future. In 1997 alone, there were numerous examples of people using encryption as an integral part of their criminal plans: 1) a convicted spy Aldrich Ames was told by the Russian Intelligence Service to encrypt computer file information that was to be passed to them; 2) an international terrorist was plotting to blow up 11 U.S.-owned commercial airlines in the Far East; his laptop computer which was seized during his arrest contained encrypted files concerning his plot; 3) a subject in a child pornography case used encryption in transmitting obscene and pornographic images of children over the Internet; and 4) a major international drug trafficking subject used a telephone encryption device to frustrate court-approved electronic surveillance. Furthermore, between 1994 and 1997, the FBI saw the number of computer-related cases utilizing encryption and/or password protection increase from 2 percent of the cases involving electronically stored information to 7 percent.
The government’s recent policy stance with respect to encryption has not been to ban encryption outright – rather, it has been to encourage the free and widespread use of strong encryption, yet to allow law enforcement officials to regulate encryption for the good of public safety by providing for a “back door” access to encrypted information in special circumstances. In March of 1998, Robert Litt, the Principal Associate Deputy Attorney General during the Clinton administration, testified before the Senate Subcommittee on the Constitution, Federalism, and Property Rights and analogized the government’s policy towards encryption to regulations governing automobiles. He argued that society “managed” the automobile, not by letting it develop completely unfettered and without regard to public safety concerns, but first by recognizing that cars could cause substantial damage to the public safety, and then by regulating the design, manufacture, and use of cars to protect the public safety. As a result, cars today are subject to numerous regulations: they must be inspected for safety on a regular basis, they are subject to minimum gasoline mileage requirements and maximum pollutant emission requirements, and they must comply with seat belt and air bag regulations. Furthermore, the laws of every jurisdiction in the United States closely regulate every aspect of driving cars on the public streets and highways, from driver’s licenses to regulation of speed to direction and flow of traffic. Litt argued that Congress and the state legislatures recognized the public safety and health threats posed by the technology of automotive transportation, even as they recognized the dramatic benefits of mobility, productivity, and industrialization that the automobile brought with it.
Litt continued with his automobile analogy by stating that the most relevant example to the policy issues posed by encryption is the practice of requiring cars to be registered and to bear license plates. More recently, federal law has required all vehicles to bear a vehicle identification number (VIN). According to Litt, America now recognize that license plates and VINs afford victims of accidents, victims of car theft, and law enforcement officials with an essential means of identifying vehicles and obtaining information on the movements of criminals. And just as legislatures in the early 1900s acted to manage the risks posed by automotive technology, the American government today urges that there be the same sensitivity to the need to preserve and advance public safety in the face of encryption in the information age. And according to Litt, such a regulatory scheme, if properly constructed, will have benefits for businesses and consumers, just like license plates.
Reflecting these concerns addressed by Robert Litt on the need to have regulatory oversight of encryption, the government has, in the past 8 years, come up with numerous proposals that would enhance law enforcement officials’ ability to access encrypted information. In 1993, the Clinton administration announced its “Clipper Chip” proposal. Under this proposal, a government-designed encryption chip, called the Clipper Chip, would become the industry standard for encryption. Everyone who wanted to encrypt and decrypt messages (i.e., their email messages or telephone conversations) would be required to use the Clipper Chip. But precisely how Clipper encrypted messages was classified. To ensure that law enforcement officials could easily tap Clipper-encrypted messages, the government would keep copies of each of the Clipper decoding keys. However, law enforcement officials’ use of these keys were restricted – only with an equivalent of a search warrant would the government be allowed to use the keys to decrypt the messages. With the Clipper Chip proposal, the Clinton administration thought that they had struck a proper balance between allowing the use of strong cryptography and answering law enforcement concerns for decrypting messages in certain circumstances. Now, individuals were allowed to use the highest level of encryption, yet government still had a back door key to each encrypted message in case they needed it.
The Clinton administration sorely miscalculated public sentiment. Rather than embrace the Clipper Chip proposal, the public lashed out, denouncing it as undermining privacy rights and being reminiscent of “an Orwellian Big Brother government.” The American Civil Liberties Union (ACLU) was one of the strongest voices against the proposal. It analogized the Clipper Chip proposal to “the equivalent of the government requiring all homebuilders to embed microphones in the walls of home and apartments.” A Time/CNN poll conducted soon after the Clipper Chip was proposed found that 80 percent of the public opposed it on the grounds that it made private information too readily accessible to government officials. The Clinton administration quickly withdrew the proposal.
Shortly after withdrawing the Clipper Chip proposal, the Clinton administration came up with a second, less-intrusive proposal. Rather than having individuals install a government-manufactured chip in order to encrypt and decrypt information, this second proposal allowed individuals to use freely their own encryption programs. In exchange, anyone who used greater than a 40-bit encryption program would be required to leave their decrypting key with a government-approved “escrow agent.” As with the original Clipper Chip proposal, law enforcement officials would be allowed access to the keys only with a court-approved warrant. However, this second proposal was also met with tremendous public opposition. The fact that the escrow agents were pre-screened and approved by government officials led the public to believe that this second proposal would have the same practical effect as the original Clipper Chip proposal in terms of potential government intrusion on individuals’ privacy rights. Public sentiment against this second proposal forced the Clinton administration to abandon it in mid-1994.
The Clinton administration’s third proposal to allow government officials to have a “back door” key to encrypted messages did not differ significantly from its second proposal. This proposal would still have used escrow agents to hold onto keys of greater than 40-bit encrypted messages; however, rather than being government approved agents, the escrow agents would have been “trusted third parties,” chosen from within the private sector. This did little to silence the critics – and by early-1995, this third proposal was also abandoned.
Finally, in 1996, the Clinton administration revealed a fourth plan and made some important changes in the direction of its policies. There would henceforth be no restrictions on the use of cryptographic systems – based on key length or technology – if those systems contained so-called “key recovery” features. That is, if encryption users could demonstrate a viable plan in which trusted third parties (possibly including “self-escrow” within user organizations) would hold (and supply government when presented “appropriate legal authority”, i.e., if government obtained a legal warrant) information that would permit recovery of code keys and decryption of data, unrestricted use of such encryption systems would be allowed. Although some in the U.S. business community continued to objected, initial reaction was much more favorable than with previous cryptography initiatives. The government had worked with U.S. business in developing the new initiative, and a number of major U.S. computer and software companies voiced support for the general principles outlined in the initiative. Others took a wait-and-see approach.
The wait was not a long one. Within months, a number of the proposal’s initial supporters had publicly or privately defected as the details of the implementation were revealed. One major sticking point was the government’s apparent desire to involve itself in frequent and detailed reviews of proprietary company business plans and progress in developing key recovery systems, as a condition for continued approval of high-level encryption systems.
As the debates over encryption policy raged on during 1996 and 1997, three major legislative proposals emerged in the 105th Congress (1997-1998). The first was called the Security and Freedom Through Encryption Act (SAFE). This act, championed by Representatives Bob Goodlatte (R-VA) and Zoe Lofgren (D-CA), was designed to promote privacy, security, and competitiveness in the Information Age. This proposal purported to: 1) affirm Americans’ freedom to use the strongest possible encryption; 2) defeat attempts to force Americans to provide the government with some government-approved third party with keys to their encrypted information; and 3) allow the U.S. to compete in the rapidly growing market for strong encryption products. U.S. businesses, as well as groups such as the ACLU and Americans for Computer Privacy, strongly supported this act.
A second proposal was called the E-Privacy Act, introduced on May 12, 1998 by Senators John Ashcroft (R-MO) [who is now, of course, the Attorney General of the United States], Patrick Leahy (D-VT), and Conrad Burns (R-MT). The bill sought to encourage the widespread availability of strong encryption without back-doors for government, and provide security for individuals’ communications, business transactions, medical records, tax returns, and other sensitive information. This bill was also backed by the ACLU and Americans for Computer Privacy.
The third proposal was called the Secure Public Networks Act, sponsored by Senators John McCain (R-AZ), Bob Kerrey (D-NE), John Kerry (D-MA), and Ernest Hollings (D-SC). Unlike the Security and Freedom Through Encryption Act, S. 909 would implement a key-recovery system, whereby Americans would be required to use government-approved third party agents to hold a spare copy of their encryption keys. The purpose of the bill was to promote national security by ensuring that law enforcement officials would have an adequate back door entry to encrypted messages that otherwise would be unbreakable. Law enforcement offices, namely the DOJ and FBI, supported this bill.
Debates over these three pieces of legislation continued for over two years. Finally, on January 12, 2000, amid political pressure from U.S. computer industry interest groups, the Clinton administration announced that encryption laws were to be liberalized. These new liberalized laws adopted much of what had been proposed in the Security and Freedom Through Encryption Act.
V. Arguments Against Restrictions on Cryptography – Free Commerce and the Right to Privacy
Opposition to restrictions on the free use of high-level encryption has come from many corners. The U.S. computer industry, led by its trade association called the Americans for Computer Privacy (ACP), and civil liberties groups, such as the ACLU, have been the most vocal opponents to the government’s proposed restrictions.
The Americans for Computer Privacy, which is supported by generous funding from big corporations like Microsoft, Intel and Cisco, argue that strong, unfettered cryptography is necessary not only for the protection of individual rights and liberties (arguments that will be discussed more in the context of the ACLU’s position), but is also important for commercial reasons. This commerce argument has two components: 1) restrictions on encryption will hold American businesses from realizing the full potential of Internet commerce within the United States; and 2) those restrictions will detriment the competitive advantage that American businesses currently enjoy over foreign competitors. With respect to the first line of argument, ACP points out that the benefits of moving business onto the Internet is undisputed – transaction costs get cut down, the speed of commerce will accelerate exponentially, and potentials for human error are reduced as transactions are digitized. However, without the security of strong encryption, customers will be wary of conducting business online. Without assurance that no one, including the government, can intercept and misuse the data that they electronically submit, ACP argues that customers will be discouraged from sending sensitive information (i.e., credit card information) over the Internet. This reluctance to use the Internet as a commerce medium will thus lead to the under-utilization of the Internet’s true business potential. Therefore, ACP believes that strong encryption without allowing for government to have a back door access to encrypted messages is indispensable for tapping the Internet’s full commerce potential.
Along with enabling companies to unlock the Internet’s full business potential within the United States, ACP argues that strong encryption without providing for a government back door is necessary to maintain U.S. companies’ strength relative to foreign companies. Currently, American companies are world leaders in computers and communications, where success in global markets is an essential ingredient in maintaining a competitive advantage. However, the market for information technology is one in which capable foreign competitors stand ready to pick up the baton of technological leadership should American companies stumble. If, for example, the United States government imposes restrictions on high-level encryption that other countries (such as France) do not, wouldn’t a customer who is greatly concerned with information security but otherwise indifferent between the two firms, want to choose to transact business with the French company? ACP fears that if this selection away from the United States happens on a macro-level, American companies will lose the competitive advantage that it currently enjoys over foreign companies.
A more prevalent line of attack against restrictions on encryption programs has been the argument that giving government a back door access to high-level encryption programs, as proposed originally by the Clinton administration and subsequently by the Secure Public Networks Act (sponsored by Senators McCain, Kerrey, Kerry, and Hollings), is an invasion of individuals’ right of privacy. This argument has been most fervently voiced by the ACLU. The ACLU warns that:
Without the right to strong, non-key recovery encryption, the black strips on the back of our credit, cash, and identity cards, the electronic keys being distributed by gasoline companies to enable the purchase of gas with the wave of a wand, the E-Z passes for paying tolls electronically, and the imminent arrival of compact digital cell phones that also function as computers, e-mailers, and pagers, will all be vulnerable to both governmental and non-governmental spying, both authorized and unauthorized.
Privacy advocates, like the ACLU generally believe that electronic surveillance, whether through bugging devices, wiretaps, or ready access to encryption keys, is fundamentally at odds with personal privacy. They argue that electronic surveillance is the worst form of a general search (prohibited by the Fourth Amendment), which necessarily captures not only the communications of its specific targets, but those of countless others who happen to come in contact with the targets or use the same lines. In their opinion, free citizens must have the ability to conduct direct, instantaneous, spontaneous, and private communication using whatever technology is available. Without the knowledge and assurance that private communications are indeed private, habits based upon fear and insecurity will gradually replace the habits of freedom.
Privacy advocates couch much of their rhetoric in terms of the provisions and purposes of the Fourth Amendment. The Fourth Amendment, which states, in pertinent part, that “The right of the people to be secure in their persons, papers, and effects, against unreasonable searches and seizures, shall not be violated…” was adopted in response to the English Parliament’s practice of giving colonial revenue officers complete discretion to search for smuggled goods by means of writs of assistance. The writs permitted colonial authorities, including British troops, to enter homes and offices at will and search any person or place they wanted. The early Americans rebelled against these general searches, and on the eve of the Declaration of Independence, Samuel Adams said that he regarded the opposition to general searches as “the Commencement of the Controversy between Great Britain and America.” It is fair to say that absolute protection from general government searches is one of America’s founding principles. [For more background on the Fourth Amendment, see Module IV.]
Privacy advocates point out that when the framers struck the original balance between personal privacy and the needs of law enforcement, remote listening devices had not yet been invented. But, they argue, had they existed, the framers would not have approved of them. Privacy advocates argue that electronic surveillance constitutes a general search by definition, not a search limited to specific objects, people, and places as required by the Fourth Amendment. Instead, wiretapping, bugs, and keys to encrypted messages intrude on the most intimate aspects of human life, hearing and seeing everything and everyone. A tap on the phone of one person necessarily captures the conversations of anyone who happens to use that phone or call that number. Likewise, unlocking one person’s encryption code subjects all people who electronically communicate with that person to government surveillance. Furthermore, privacy advocates argue that the requirement that law enforcement obtain a warrant before unlocking a person’s encryption code will not ameliorate the problem. Electronic eavesdropping cannot be regulated by a warrant precisely because of its dragnet quality; the object to be seized or the premises to be searched cannot be limited or even specified, because it is in the very nature of the technology to capture everything. Moreover, the warrant requirement will not, in all likelihood, serve as an adequate safeguard to obtaining encryption keys. With respect to wiretaps, the ACLU points out that although government agents must obtain a warrant, their requests are almost never turned down by judges or magistrates – in fact, between 1990 and 1998, only one request by a law enforcement for a wiretap was rejected.
Not only will the existence of a government back door to encrypted messages have a chilling effect on private electronic communication, privacy advocates argue that non-key recovery systems (i.e., unbreakable encryption programs) are not detrimental to law enforcement, as the DOJ and FBI fear. The ACLU argues that the government’s own records show that electronic surveillance is of marginal utility in preventing or solving serious crimes. Between 1987 and 1998, fewer than 0.2 percent of all law enforcement wiretap requests were made in the investigation of bombings, arsons, or firearms. Nor is wiretapping often used in other crimes of violence, such as homicide, assault, rape, robbery, and burglary. Instead, the vast majority of wiretaps and other forms of surveillance have been authorized in connection with vice crimes, such as gambling and drug offenses – in fact, 83 percent of the wiretap cases between 1987 and 1998 were crimes of this nature. Although vice crimes are nonetheless punishable felonies, they do not rise to the same level of endangering public safety as crimes of violence. Privacy advocates believe that the disproportionate use of electronic surveillance to vice crimes as opposed to crimes of violence would be similar with respect to government’s access to encrypted programs. If true, privacy advocates argue that having unbreakable encryption would not be detrimental to law enforcement.
In contrast to the marginal utility that electronic surveillance has to the protection of public safety, privacy advocates argue that such surveillance has (and will continue if the government is allowed a back door to encryption programs) resulted in demonstrable violations of the privacy rights of vast numbers of Americans. According to statistics released by the Administrative Office of the U.S. Courts and the DOJ, 2.2 million conversations were captured in 1996, of which a total of 1.7 million conversations were deemed not incriminating by prosecutors.
Lastly, privacy advocates argue that high-level encryption provides individuals with a self-help mechanism against computer crime, theft, and fraud. As discussed in earlier section, encryption prevents sensitive information, such as bank records, medical information, and credit card numbers, from being misused by hackers and other Internet snoopers. Therefore, high-level, unbreakable encryption can actually deter and prevent crime. According to privacy advocates, by requiring back door access to all encrypted messages via a key-recovery system, the government will be introducing a flaw into the effective self-help mechanism that encryption provides. By its very design, a key recovery system introduces a system weakness; it is deliberately designed to allow access in certain exceptional circumstances. Therefore, the fear is that if the procedures that protect against abuse of that access somehow failed, information would be left unprotected. Since the government, under a key recovery system, would have the keys to all high-level encrypted messages, what would happen if the keys got into the wrong hands? What would happen if corrupt police officers or hackers that break into government files were able to access these keys? The potential for crime and danger in this scenario would be endless.
Furthermore, a key recovery system could weaken the confidentiality provided by an encryption system by providing an access path that can be compromised. For example, if a party external to a corporation has the encryption keys to that corporation’s encrypted information in escrow, the corporation is more vulnerable to a loss of confidentiality, because the external party can become the target of theft, extortion, or blackmail by unauthorized parties who are seeking that information. As a result, privacy advocates argue that, contrary to what the FBI and DOJ believe, there would actually be more crime if society operated under a key recovery system than under a non-key recovery system.
The economic analysis and constitutional rhetoric put forth by the opponents to a government key recovery system has resonated with the American public and with both the House of Representatives and the Senate. As a result of this public support, as well as a strong lobbying campaign put forth by the American computer industry prior to the 2000 Presidential election, the Clinton administration backed off from its key recovery system proposal and greatly liberalized the government’s policy with respect to high-level encryption. As it stands now, government does not have a back door access to high-level encrypted information and, with some minor exceptions, encryption has now been left unfettered by government regulation. Although the debate of how encryption policy should be formulated has come to rest for the time being, is having strong, unfettered encryption necessarily good for America and its citizens and consumers?
V. Another Way to Look at the Encryption Policy Debate: Individuals vs. Marketeers
The January 12, 2000 decision made by the Clinton administration to liberalize encryption laws was hailed as a victory by privacy and civil liberty interest groups as well as by the U.S. computer industry. ACP Co-Chairman, Jack Quinn, praised the decision, stating that:
ACP is extremely gratified by the new encryption regulations. They are more in step with the economic realities of the Information Age, while protecting our nation’s vital security and law enforcement needs. And, they strike a balance between security and America’s commercial interests.
According to both the ACLU and the ACP, the leading opponents of the government-proposed key recovery system, having strong, unfettered encryption would allow individuals to protect themselves: 1) against computer crimes committed by other individuals; and 2) against government’s invasion of their privacy. Classified in general terms, the ACLU and ACP believe that a policy that promotes strong, unfettered encryption will help protect individuals vis-à-vis other individuals as well help protect individuals vis-à-vis the state. Most, if not all, of the encryption debate has revolved around these two power relationships. The debates, however, have not addressed a third relationship that encryption will affect. This relationship concerns the power dynamic between the individuals as consumers and online vendors (henceforth referred to as marketeers). How would a policy that promotes strong, unfettered encryption affect the relationship between individual consumers and marketeers?
According to Stanford Law School Professor, Lawrence Lessig, “encryption technologies are the most important technological breakthrough in the last one thousand years.” Although Lessig himself concedes that his statement may be a “slight exaggeration,” his assertion may not be too far off base. Encryption, as you now know, can be used to encode data in such a way that it is very difficult or impossible to decrypt by unauthorized people. Encryption thus enables a person to control data in certain way – i.e., by restricting who can access and read his message. Controlling who can access and read an electronic communication is but one way that encryption can be used as a means to control data. With encryption, a person can control precisely what the recipient can read, how long the recipient has to read the message, and whether or not the recipient can store the message for future use. Such control can be applied to files and programs other than just documents or text files (i.e., sound files, graphics, digital videos, and application programs). Furthermore, encryption can be used in a way that will enable the sender of the encrypted information to monitor and track the use of that information. Through this monitoring and tracking, the sender may be able to easily gather personal information about the recipient or his habits without the recipient’s knowledge that he is being tracked. With almost every facet of the economy becoming more computerized, with our society truly becoming a digital society, the control over data that encryption provides is an incredible power. Viewed in this way, perhaps Lessig’s statement is not as much of an exaggeration as it may seem at first glance.
One group that would benefit from encryption is marketeers. Marketeers, such as Amazon.com, can profit in three principle ways. First and most obviously, they make money by selling the products that they advertise on their website. For Amazon.com, that would mean selling their books to Internet users who shop online. Second, marketeers make money through advertisements posted on their website. If you visit Amazon.com or any other online vendor, you are likely to see numerous advertisements that promote merchandise and services of other companies. Those other companies, naturally, pay websites like Amazon.com for that website space, in much the same way that a company would pay a newspaper for publishing their advertisement. Third, and most relevant to encryption, marketeers can make money by selling information that they collect about their website users to other companies. Alternatively (and now more frequently), sites like Amazon or Yahoo can mine the information they collect and market their databases to third parties while keeping control of the information. How does this work? Keep in mind that in the digital age (and the Internet, in particular), one of the most valuable assets is information. Amazon.com can thus sell the information that they gather about their customers to other companies or data marketing firms who want this information just as they would sell any other commodity or asset. Most of the information that Amazon.com and other online vendors gather about their members are obtained during a person’s very first visit to their website. New users to a website such as Amazon.com are required to fill out an informational survey in order to become a member to the site. Because membership is required to make purchases, anyone who wants to use Amazon.com to buy books (or any other product) must provide the requested information.
The information that Amazon.com requires to become a member of their website may seem harmless--name, mailing address, email address, age (or age range), gender, hobbies/interests, and occupation. To many companies and data marketing firms, however, this information is very valuable, particularly when linked to a user’s reading habits. Borrowing Amazon’s database would grant a marketeer access to enormous interior information about a customer. With access to such a database, it would be very easy for you, as (e.g.) Sports Illustrated magazine’s chief of advertising, to market to the 20-25 males sports fanatics who have read reviews of at least five sports related books, and who also read about wine and law. And by gathering many marketeer databases, you would be able to profile large numbers of people who fit that demographic. The ability to target your advertising to that demographic would thus be realistic.
Access to marketeers’ member databases is very valuable to all sorts of companies who want to target their advertising campaigns. As a result, marketeers can generate significant fees by selling marketing information from their databases to other companies. And furthermore, the more information that marketeers can extract and gather from its customers, the more information they can sell to companies who want to buy access to their databases. Therefore, marketeers are constantly looking for ways that they can accumulate more information about their customers’ behavior and preferences.
Encryption provides a structural mechanism by which marketeers can extract ever more detailed information about their customers. Larry Lessig demonstrates how encryption can be utilized not only to restrict access but also to monitor and control usage. Recall that in general, encryption is a mechanism by which a person who sends data can control that data. One of the controls encryption allows is monitoring and tracking of data, including downloaded data. Therefore, the sender can encrypt a piece of information in such a way that he can monitor and track how the recipient uses the information, even to determine the time of day downloaded data might be read or utilized. In the Internet context, encryption enables control of information that would have been unthinkable in an earlier era. It is at least theoretically possible for every bit of downloaded data to carry markers that allow the “seller” of the information to control what information is viewed by what persons, for what period of time, in what context. If the encryption key is housed in an interlinked environment, it is a simple matter for the description of the actual use then to be sent back to the originator. Encryption would enable marketeers such as Amazon.com to encrypt all the information on their website and in downloaded data (such as e-books) in a way that would enable it to gather all sorts of information about members’ online and offline reading habits.
Whereas marketeers who sell database information and the companies that purchase this information stand to benefit from unregulated encryption, individual consumers who use Amazon.com or other marketeers’ websites stand to lose via invasion of their privacy rights. The privacy concerns raised by the spector of an Amazon.com tracking every mouseclick on its website pales in comparison to the encryption-driven potential of monitoring all information that someone reads. Such use of encryption would seem to raise the same privacy concerns that the ACLU and ACP addressed with respect to the government having a “back door” key to all encrypted information. Isn’t the concept of having a “Big Brother” marketeer just as frightening in terms of privacy concerns as having a “Big Brother” government?
If the privacy concern has still not been made clear, consider another example. Suppose that you decide to spend an afternoon at a library or bookstore, perhaps to buy a book, but mainly just to browse and leaf through various books and magazines. Now consider that instead of being able to freely move about the library or bookstore and look at any book without others knowing it, the librarian or store owner follows you around with pen and paper, writing down the title and author of every book that you look at. If you stop for a moment to peruse pages 3, 6-9, 41, and 101-115, the owner keeps track of the pages as well as the amount of time spent on each page. Furthermore, consider that the librarian or bookstore owner takes down your name and email address and markets the list of books that you looked at to various companies and data marketing firms, who subsequently inundate you with advertisements to purchase their products. How would you feel? Not only would you feel a bit annoyed at the inconvenience of receiving unsolicited advertisements, but you might also believe that this whole scheme violates your fundamental privacy rights. This is exactly the situation that strong, unfettered encryption has the potential create in the context of marketeers and online consumers.
The ACLU and ACP have advocated strongly for unfettered encryption, arguing that it will lead to greater protection of individual privacy. But is unregulated encryption a panacea for privacy concerns? It may be true that unregulated encryption protects individuals’ privacy interests vis-à-vis other individuals, as well as their privacy interests vis-à-vis the state. However, as discussed above, a regime with unregulated encryption may actually be a cause for a greater invasion of individuals’ privacy – vis-à-vis marketeers and companies that purchase information databases from marketeers.
Despite the Clinton administration’s decision to liberalize encryption restrictions, debates will continue over whether or not encryption should be regulated and how much it should be regulated. As concerns for individual privacy are weighed during these debates, we think that one of the important issues that should be considered is the effect that unrestricted encryption could have on the individual consumer’s privacy interests vis-à-vis marketeers. Viewed from this perspective, it is possible that the arguments associated with privacy concerns are not as clear and one-sided as the ACLU and ACP make it out to be.
We do not offer this viewpoint as a means to rebut the ACLU and ACP’s argument and to advocate on behalf of the government’s position. Instead, we mean to present a viewpoint that deserves to be considered when policymakers sit down and debate the merits of having restrictions on encryption. Only with an understanding of the full impact of encryption technology can a meaningful policy on encryption restrictions be formulated.
 This paper was originally prepared by Teddy Kang (HLS ’01) as a description of the background of encryption as related to cyberspace. Section V (“Another Way to Look at the Encryption Policy Debate: Individuals v. Marketeers”) has been substantially altered to focus on potential future risks to privacy posed by encryption itself.
 Such as a hacker or a snooper.
 How does this interception work? Recall that after typing up the “HELLO” message to Bill and clicking the “send” button to transmit the message, the TCP/IP software on your computer breaks down the message into datagrams. These datagrams then get directed by routers from your LAN to Bill’s LAN via regional networks and, if necessary, backbones. However, the routing process is not completely secure. Therefore, hackers, expert programmers who find special tricks for getting around computer security, or other Internet snoopers, have the ability to intercept the datagrams as they get sent over the Internet public wires. By intercepting these datagrams, they can easily read your email message.
 Lawrence Lessig, Code and Other Laws of Cyberspace 36 (1999).
 See id.
 Ronald D. Lee, Associate Deputy Attorney General of the Department of Justice, Statement Before the House Committee on the Judiciary, Subcommittee on Courts and Intellectual Property (March 4, 1999).
 See Dan Froomkin, Deciphering Encryption, Washington Post, May 8, 1998, at A4.
 See Lee, supra note 12.
 See Louis J. Freeh, Director of the Federal of Investigations, Statement Before the Permanent Select Committee on Intelligence, United States House of Representatives (Sep. 9, 1997).
 See id.
 See id.
 See id.
 See id.
 See Robert S. Litt, Principal Associate Deputy Attorney General, Testimony Before the Subcommittee on the Constitution, Federalism, and Property Rights Committee, United States Senate (Mar. 17, 1998).
 See id.
 See id.
 See id.
 It was the VIN in the Oklahoma City bombing case that led the FBI to the truck rental office at which Timothy McVeigh rented the truck he used.
 Litt, supra note 25.
 See id.
 See Kenneth Flamm, Deciphering the Cryptography Debate, Brookings Institute Policy Debate #21, July 1997.
 See id.
 See id.
 See id.
 Froomkin, supra note 15.
 See id.
 See id.
 See id.
 Flamm, supra note 32.
 See id.
 See Roberto Suro and Elizabeth Corcoran, U.S. Law Enforcement Wants Keys to High-Tech Cover, Washington Post, March 30, 1998, at A4.
 See id.
 A system that enabled recovery of their own encrypted business data, in fact, was actually useful to companies in dealing with the risks of employee turnover.
 Suro and Corcoran, supra note 44.
 See id.
 H.R. 695, 105th Cong. (1997).
 S. 2067, 105th Cong. (1998).
 S. 909, 105th Cong. (1998).
 David E. Sanger and Jeri Clausing, U.S. Removes More Limits On Encryption, New York Times, Jan. 13, 2000, at D1.
 U.S. Policy on Encryption (visited Apr. 4, 2001) <http://www.computerprivacy.org/about>.
 See Flamm, supra note 32.
 See id.
 See id.
 See id.
 See id.
 Big Brother in the Wires: Wiretapping in the Digital Age, ACLU Special Report (Mar. 1998).
 See id.
 See id.
 See id.
 See id.
 See id.
 Kenneth W. Dam and Herbert S. Lin, Cryptography’s Role in Securing the Information Society 181 (1996).
 See id at 182.
 Sanger and Clausing, supra note 55.
 See id.
 ACP Extremely Gratified by Modernized Encryption Policy (posted Jan. 12, 2000) <http://www.computerprivacy.org/news/947779537.shtml>.
 Lessig, supra note 4 at 35.
 Id. at 36.
 See id.
 Yahoo recently announced that it will start sending users of its site e-mail marketing messages on behalf of its own services, even if its users had previously requested not to receive any marketing from Yahoo.
 A recent New York Times article describes just how valuable even limited information is:
“Direct Media, a mailing list broker in Greenwich, Conn., offers access to 2.9 million Lycos users at a cost of $125 per thousand names for a single mailing. (An extra $15 per thousand lets marketers select users showing an interest in a topic like cats or gambling.) Advertisers typically pay for the right to send a single mailing or make a single phone call to a name on a list they rent; they do not own the information outright.”
 See Lessig, supra note 4 at 35.
 See id.
 Sanger and Clausing, supra note 55.