CRYPTO WARS: Thu, 22 Apr 93 10:53 PDT: "Thoughts on Clipper" Had really hoped I wouldn't have to drag this out again more than 20 years later ...
- - - Date: Thu, 22 Apr 93 10:53 PDT From: [email protected] (Lauren Weinstein; PRIVACY Forum Moderator) Subject: Thoughts on Clipper Greetings. Well, as you might imagine, considerable mail has been coming into PRIVACY Forum regarding the Clipper chip announcement and related materials. A number of persons have asked me for my thoughts regarding the announcement and its possible implications for privacy-related issues. What follows is purely my individual, personal opinion. I apologize in advance for its length. I'm not a cryptography expert. I'm not the right person to judge the algorithmic strengths or weaknesses of a given encryption system. It is possible that a "team of experts" may determine that the Clipper chip is a powerful encryption mechanism, free (as far as can be determined) of significant algorithmic weaknesses which would impact its usefulness for the classes of applications and level of security for which it is designed. It could be argued that proper analysis of such a cryptographic system could only be conducted in an "open" environment, and that many cryptographic experts may be unwilling to participate in a "classified" analysis either out of general principles or concerns about possible impacts on their future ability to publish other cryptographic works. However, those are not the points that directly impact privacy, and I'll leave them for others to explore. For the sake of the argument, let's assume Clipper uses a strong algorithm, free from surprise "back-doors" or other weaknesses. The privacy concerns regarding Clipper focus directly on its "key-escrow" aspects and the implications of a cryptographic system specifically designed to allow access by third parties to "private" communications. If Clipper remains purely a choice to be freely accepted or rejected by the marketplace, and individual users, the primary question revolves around whether or not potential users of the system are fully informed regarding all aspects of the system (including both the relative strength of the algorithm itself and the existence of the key-escrow mechanism) and choose to use the system after being so informed. To the extent that manufacturers choose to use the system in mass-market products, the ability of users to select other systems may be impacted, but that's an issue that market forces can help rectify. There's a more ominous potential problem, however. Reading "between the lines" of the Clipper announcement, at least the possibility of future legislation being proposed to ban powerful non-key-escrow systems seems quite real. If such legislation became law, the negative impact on personal privacy rights could be serious. If such legislation were combined with the passage of concepts embodied in the FBI Digital Telephony Proposal, the potential negative impact could be quite dramatic. Would people put up with being told that from now on they had to send copies of their house keys to outside entities who would hold them in case they were needed by law enforcement? Pretty doubtful. Or would they go along with the idea that cameras would be mounted throughout their homes but would only be activated under court order? Not very likely. We're permitted to speak to each other in person using whatever languages (or codes, presumably) that we like--even though this could presumably limit the effectiveness of court-ordered bugs on a premises. We can write letters (or send faxes) using whatever symbolic systems we might wish. Is there something significantly different about a phone call which should subject it to completely different privacy rules? That's a core question in this debate. Properly authorized wiretaps can of course have significant law enforcement value. But law enforcement using authorized surveillance techniques is one thing; requiring that technology be designed in such a way as to make surveillance simpler (or for that matter, possible in a given situation) is something entirely different. As technology evolves, techniques of law enforcement, like those of war, are subject to change. Centuries ago, a suit of full body armor was protection against most of the weapons of the era. With the introduction of gunpowder in Europe, armor rapidly became obsolete--upsetting both the governments who had invested heavily in armor and those who made the armor! This is an imperfect analogy of course, but the point is that no particular technology can be expected to be equally useful forever, and that attempts to "hold back" other aspects of technology to "protect" the usefulness of another part are wasteful and usually doomed to ultimate failure. This is especially true now with computer-oriented technologies, where software-based techniques of immense power are within the grasp of anyone with a personal computer. The genie of powerful software-based encryption systems is out of the bottle--it's doubtful that he can be successfully forced back in. Of course, the more convenient a particular encryption technology is to use, the more likely it is to get used. Theoretically "perfect" systems have long existed (such as one-time pads) but they have generally been relatively inconvenient to manage logistically (modern computer technology, however, is changing even this). Does concern about key-escrow systems imply a distrust of those organizations or agencies who would be currently holding those keys? No, not necessarily. But organizations and governments change. The same entities you trust at one time may abuse their powers later. Given that access to a Clipper key-escrow pair would give the ability to decode all future and all previously recorded past transmissions between affected units, we're talking about a power that must at least be carefully scrutinized before being handed over. Perhaps this is all largely academic. Maybe there's not going to be any push for the limitation or banning of powerful non-key-escrow encryption systems. In that case, this has all been an interesting intellectual exercise. But since the crystal ball is unclear, it is to our benefits to carefully consider all the aspects of these technologies now, before they become widely entrenched. One final point. I've heard considerable numbers of persons referring to Clipper as a "Clinton" plan. I think it's fairly clear that Clipper must have been in development for years, going back for at least one administration and perhaps even farther. It seems likely that the Clinton administration was presented with an essentially complete program, and made the announcement based on the briefings and information made available to them at that time. Since the project apparently proceeded with only very limited input from the outside cryptography or computer science communities, it seems likely that the Clinton administration may not have heard a range of viewpoints regarding the technology and its implications before the announcement. The press release announcing Clipper said that there is a desire to enter into a dialogue regarding these technologies. So if you have opinions about Clipper or similar systems, either positive or negative, I'd urge you to make them known, in writing, to the administration and to other elected representatives as you see fit. - - - --Lauren-- Lauren Weinstein ([email protected]): http://www.vortex.com/lauren Founder: - Network Neutrality Squad: http://www.nnsquad.org - PRIVACY Forum: http://www.vortex.com/privacy-info Co-Founder: People For Internet Responsibility: http://www.pfir.org/pfir-info Member: ACM Committee on Computers and Public Policy I am a consultant to Google -- I speak only for myself, not for them. Lauren's Blog: http://lauren.vortex.com Google+: http://google.com/+LaurenWeinstein Twitter: http://twitter.com/laurenweinstein Tel: +1 (818) 225-2800 / Skype: vortex.com _______________________________________________ pfir mailing list http://lists.pfir.org/mailman/listinfo/pfir
