from the going-back dept
Mike Godwin (you know who he is) was recently going through some of his earlier writings, and came across an essay (really an outline) he had written to the Cypherpunks email list 25 years ago, in April of 1993 concerning the Clipper Chip and early battles on encryption and civil liberties. If you don’t recall, the Clipper Chip was an early attempt by the Clinton administration to establish a form of backdoored encryption, using a key escrow system. What became quite clear in reading through this 25-year-old email is just how little has changed in the past 25 years. As we are in the midst of a new crypto war, Godwin has suggested republishing this essay from so long ago to take a look back at what was said back then and compare it to today.
From: Mike Godwin
Subject: Some thoughts on Clipper and the Constitution To: e*c Date: Mon, 26 Apr 93 11:15:17 EDT
Note: These notes were a response to a question during Saturday’s Cypherpunks meeting about the possible implications of the Clipper Chip initiative on Fourth Amendment rights. Forward to anyone else who might think these interesting.
Notes on Cryptography, Digital Telephony, and the Bill of Rights By Mike Godwin
A. The recent announcement of the federal government’s “Clipper Chip” has started me thinking again about what the principled “pure Constitutional” arguments a) opposed to Digital Telephony and b) in favor of the continuing legality of widespread powerful public-key encryption.
B. These notes do *not* include many of the complaints that have already been raised about the Clipper Chip initiative, such as:
(1) Failure of the Administration to conduct an inquiry before embracing a standard,
(2) Refusal to allow public scrutiny of the chosen encryption algorithm(s), which is the normal procedure for testing a cryptographic scheme, and
(3) Failure of the administration to address the policy questions raised by the Clipper Chip, such as whether the right balance between privacy and law-enforcement needs has been struck.
C. In other words, they do not address complaints about the federal government’s *process* in embracing the Clipper Chip system. They do, however, attempt to address some of the substantive legal and Constitutional questions raised by the Clipper Chip and Digital Telephony initiatives.
II. Hard Questions from Law Enforcement
A. In trying to clarify my own thinking about the possible Constitutional issues raised by the government’s efforts to guarantee access to public communications between individuals, I have spoken and argued with a number of individuals who are on the other side of the issues from me, including Dorothy Denning and various respresentatives of the FBI, including Alan McDonald.
B. McDonald, like Denning and other proponents both of Digital Telephony and of a standard key-escrow system for cryptography, is fond of asking hard questions: What if FBI had a wiretap authorization order and couldn’t implement it, either because it was impossible to extract the right bits from a digital-telephony data stream, or because the communication was encrypted? Doesn’t it make sense to have a law that requires the phone companies to be able to comply with a wiretap order?
C. Rather than respond to these questions, for now at least let’s ask a different question. Suppose the FBI had an authorization order for a secret microphone at a public restaurant. Now suppose it planted the bug, but couldn’t make out the conversation it was authorized to “seize” because of background noise at the restaurant. Wouldn’t it make sense to have a law requiring everyone to speak more softly in restaurants and not to clatter the dishes so much?
D. This response is not entirely facetious. The Department of Justice and the FBI have consistently insisted that they are not seeking new authority under the federal wiretap statutes (“Title III”). The same statute that was drafted to outline the authority for law enforcement to tap telephonic conversations was also drafted to outline law enforcement’s authority to capture normal spoken conversations with secret or remote microphones. (The statute was amended in the middle ’80s by the Electronic Communications Privacy Act to protect “electronic communications,” which includes e-mail, and a new chapter protecting _stored_ electronic communications was also added.)
E. Should we understand the law the way Digital Telephony proponents insist we do–as a law designed to mandate that the FBI (for example) be guaranteed access to telephonic communications? Digital Telephony supporters insist that it merely “clarifies” phone company obligations and governmental rights under Title III. If they’re right, then I think we have to understand the provisions regarding “oral communications” the same way. Which is to say, it would make perfect sense to have a law requiring that people speak quietly in public places, so as to guarantee that the government can bug an oral conversation if it needs to.
F. But of course I don’t really take Digital Telephony as an initiative to “clarify” governmental prerogatives. It seems clear to me that Digital Telephony, together with the “Clipper” initiative, prefigure a government strategy to set up an information regime that precludes truly private communications between individuals who are speaking in any way other than face-to-face. This I think is an expansion of government authority by almost any analysis.
III. Digital Telephony, Cryptography, and the Fourth Amendment
A. In talking with law enforcement representatives such as Gail Thackeray, one occasionally encounters the view that the Fourth Amendment is actually a _grant_ of a Constitutional entitlement to searches and seizures. This interpretation is jolting to those who have studied the history of the Fourth Amendment and who recognize that it was drafted as a limitation on government power, not as a grant of government power. But even if one doesn’t know the history of this amendment, one can look at its language and draw certain conclusions.
B. The Fourth Amendment reads: “The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated, and no Warrants shall issue, but upon probable cause, supported by Oath or affirmation, and particularly describing the place to be searched, and the persons or things to be seized.”
C. Conspicuously missing from the language of this amendment is any guarantee that the government, with properly obtained warrant in hand, will be _successful_ in finding the right place to be searched or persons or things to be seized. What the Fourth Amendment is about is _obtaining warrants_–similarly, what the wiretap statutes are about is _obtaining authorization_ for wiretaps and other interceptions. Neither the Fourth Amendment nor Title III nor the other protections of the ECPA constitute an law-enforcement _entitlement_ for law enforcement.
D. It follows, then, that if digital telephony or widespread encryption were to create new burdens for law enforcement, this would not, as some law-enforcement representatives have argued, constitute an “effective repeal” of Title III. What it would constitute is a change in the environment in which law enforcement, along with the rest of us, has to work. Technology often creates changes in our social environment–some, such as the original innovation of the wiretap, may aid law enforcement, while others, such as powerful public-key cryptography, pose the risk of inhibiting law enforcement. Historically, law enforcement has responded to technological change by adapting. (Indeed, the original wiretaps were an adaptation to the widespread use of the telephone.) Does it make sense for law enforcement suddenly to be able to require that the rest of society adapt to its perceived needs?
IV. Cryptography and the First Amendment
A. Increasingly, I have come to see two strong links between the use of cryptography and the First Amendment. The two links are freedom of expression and freedom of association.
B. By “freedom of expression” I mean the traditionally understood freedoms of speech and the press, as well as freedom of inquiry, which has also long been understood to be protected by the First Amendment. It is hard to see how saying or publishing something that happens to be encrypted could not be protected under the First Amendment. It would be a very poor freedom of speech indeed that dictated that we could *never* choose the form in which we speak. Even the traditional limitations on freedom of speech have never reached so far. My decision to encrypt a communication should be no more illegal than my decision to speak in code. To take one example, suppose my mother and I agree that the code “777”, when sent to me through my pager, means “I want you to call me and tell me how my grandchild is doing.” Does the FBI have a right to complain because they don’t know what “777” means? Should the FBI require pager services never to allow such codes to be used? The First Amendment, it seems to me, requires that both questions be answered “No.”
C. “Freedom of association” is a First Amendment right that was first clearly articulated in a Supreme Court case in 1958: NAACP v. Alabama ex rel. Patterson. In that case, the Court held that Alabama could not require the NAACP to disclose a list of its members residing in Alabama. The Court accepted the NAACP’s argument that disclosure of its list would lead to reprisals on its members; it held such forced disclosures, by placing an undue burden on NAACP members’ exercise of their freedoms of association and expression, effectively negate those freedoms. (It is also important to note here that the Supreme Court in effect recognized that anonymity might be closely associated with First Amendment rights.)
D. If a law guaranteeing disclosure of one’s name is sufficiently “chilling” of First Amendment rights to be unconstitutional, surely a law requiring that the government be able to read any communications is also “chilling,” not only of my right to speak, but also of my decisions on whom to speak to. Knowing that I cannot guarantee the privacy of my communications may mean that I don’t conspire to arrange any drug deals or kidnapping-murders (or that I’ll be detected if do), but it also may mean that I choose not to use this medium to speak to a loved one, or my lawyer, or to my psychiatrist, or to an outspoken political activist. Given that computer-based communications are likely to become the dominant communications medium in the next century, isn’t this chilling effect an awfully high price to pay in order to keep law enforcement from having to devise new solutions to new problems?
V. Rereading the Clipper Chip announcements
A. It is important to recognize that the Clipper Chip represents, among other things, an effort by the government to pre-empt certain criticisms. The language of announcements makes clear that the government wants us to believe it has recognized all needs and come up with a credible solution to the dilemma many believe is posed by the ubiquity of powerful cryptography.
B. Because the government is attempting to appear to take a “moderate” or “balanced” position to the issue, its initiative will tend to pre-empt criticisms of the government’s proposal on the grounds of *process* alone.
C. But there is more to complain about here than bad process. My rereading of the Clipper Chip announcements will reveal that the government hopes to develop a national policy that includes limitations on some kinds of cryptography. Take the following two statements, for example:
D. ‘We need the “Clipper Chip” and other approaches that can both provide law-abiding citizens with access to the encryption they need and prevent criminals from using it to hide their illegal activities.’
E. ‘The Administration is not saying, “since encryption threatens the public safety and effective law enforcement, we will prohibit it outright” (as some countries have effectively done); nor is the U.S. saying that “every American, as a matter of right, is entitled to an unbreakable commercial encryption product.” ‘
F. It is clear that neither Digital Telephony nor the Clipper Chip make any sense without restrictions on other kinds of encryption. Widespread powerful public-key encryption, for example, would render useless any improved wiretappability in the communications infrastructure, and would render superfluous any key-escrow scheme.
G. It follows, then, that we should anticipate, consistent with these two initiatives, an eventual effort to prevent or inhibit the use of powerful private encryption schemes in private hands.
H. Together with the Digital Telephony and Clipper Chip initiatives, this effort would, in my opinion, constitute an attempt to shift the Constitutional balance of rights and responsibilities against private entities and individuals and in favor of law enforcement. They would, in effect, create _entitlements_ for law enforcement where none existed before.
I. As my notes here suggest, these initiatives may be, in their essence, inconsistent with Constitutional guarantees of expression, association, and privacy.