Using Nanotechnology Safely

A safe personal nanofactory design must build approved products, while refusing to build unapproved products. It must also be extremely tamper-resistant; if anyone found a way to build unapproved products, they could make an unrestricted, unsafe nanofactory, and distribute copies of it. The product approval process must also be carefully designed, to maximize the benefits of the technology while minimizing the risk of misuse. Restricted nanofactories avoid the extreme risk/benefit tradeoff of other nanotechnology administration plans, but they do require competent administration.

One way to secure a personal nanofactory is to build in only a limited number of safe designs. The user could ask it to produce any one of those designs, but with no way to feed in more blueprints, the factory could never build anything else. This simple scheme is fairly reliable, but not very useful. It also poses the risk that someone could take apart the factory and find a way to reprogram its design library.

A more useful and secure scheme would be to connect the PN to a central controller, and require it to ask for permission each time it was asked to manufacture something. This would allow new designs to be added to the design library after the nanofactory was built. In addition, the PN would have to report its status back to the central controller. The system could even be designed to require a continuous connection; a factory disconnected from the network would permanently disable itself.

This would greatly reduce the opportunity to take the factory apart, since it could report the attempt in real time, and failed attempts would result in immediate arrest of the perpetrator. This permanent connection would also allow the factory to be disabled remotely if a security flaw were ever discovered in that model. Finally, a physical connection would allow the location of the factory to be known, and jurisdictional limits to be imposed on its products.

Current cryptographic techniques permit verification and encryption of communication over an unsecured link. These are used in smart cards and digital cellular phones, and will soon be used in digital rights management. Using such techniques, each personal nanofactory would be able to verify that it was in communication with the central library. Only designs from the library could be manufactured. In addition, each design could come with a set of restrictions. For example, medical tools might only be manufactured at the request of a doctor. Commercial designs could require payment from a user. Designs under development could be manufactured only by the inventor, until they were approved and released. A design that did not come from the central library would not have the proper cryptographic signature, and the factory would simply refuse to build it.