Nanofactory Technology: Regulating Risk, Preserving Benefit

It is generally assumed, incorrectly, that devices built with nanotechnology must be quite small. This has led to fears that molecular manufacturing systems will be hard to control and easy to steal. In fact, as analyzed by Drexler and others in the field, the products of nanoscale mechanochemical plants can be attached together within the enclosure of a single device. Small building blocks can be joined to make bigger blocks; these blocks can be joined with others, and so on to form a product. This process is called convergent assembly, and it allows the creation of large products from nanoscale parts. In particular, convergent assembly will allow one nanofactory to build another nanofactory. There is no need to use trillions of free-floating assembler robots; instead, the assemblers—now called fabricators—are securely fastened inside the factory device, where they feed the smallest conveyor belts.

A typical personal nanofactory (PN) might be the size of a microwave oven. Since the fabricators are fastened into the factory and dependent on its power grid, they have no need to navigate around the product they are building—this improves efficiency—and they have no chance of functioning independently. In addition, the entire nanofactory can be controlled through a single interface, which allows restrictions to be built into the interface. It can simply refuse to produce any product that has not been approved. (The improved security of tethered nanotechnology factories has been a theme in at least one work of science fiction.)

If a PN will only build safe products, and will refuse to build any product that has not been approved as safe, then the factory itself can be considered safe. It could even build a duplicate PN on request. With the restrictions built in, the second one would be as safe as the first. As long as the restrictions work as planned, there is no risk of grey goo, no risk of undesirable weapons or unapproved products, and no risk of producing unrestricted nanofactories that could be used to make bad products.

At the same time, products that were approved could be produced in any quantity desired. The products could even be customized, within limits—and the limits could be quite broad, for some kinds of products. If desired, the PNs (and the products) could have tracking devices built in to further deter inappropriate use.

With personal nanofactories that can only produce approved designs, the safety of molecular manufacturing does not depend on restricting the use of the factories. Instead, it depends on choosing correctly which products to approve. The nanofactory itself, as a product, can be approved for unlimited copying. This means that the abundant, cheap, and convenient production capability of advanced nanotechnology can be achieved without the risks associated with uncontrolled molecular manufacturing. A two-dimensional view of the risks of nanotechnology, which separates the means of production from the products, allows the design and implementation of policy that is minimally restrictive, yet still safe.