Advanced Nanotechnology And Its Risks

As early as 1959, Richard Feynman proposed building devices with each atom precisely placed. In 1986, Eric Drexler published an influential book, Engines of Creation, in which he described some of the benefits and risks of such a capability. If molecules and devices can be manufactured by joining individual atoms under computer control, it will be possible to build structures out of diamond, 100 times as strong as steel; to build computers smaller than a bacterium; and to build assemblers and mini-factories of various sizes, capable of making complex products and even of duplicating themselves.

Drexler's subsequent book, Nanosystems, substantiated these remarkable claims, and added still more. A self-contained tabletop factory could produce its duplicate in one hour. Devices with moving parts could be incredibly efficient. Molecular manufacturing operations could be carried out with failure rates less than one in a quadrillion. A computer would require a miniscule fraction of a watt and one trillion of them could fit into a cubic centimeter. Nanotechnology-built fractal plumbing would be able to cool the resulting 10,000 watts of waste heat. It seems clear that if advanced nanotechnology is ever developed, its products will be incredibly powerful.

As soon as molecular manufacturing was proposed, risks associated with it began to be identified. Engines of Creation described one hazard now considered unlikely, but still possible: grey goo. A small nanomachine capable of replication could in theory copy itself too many times. If it were capable of surviving outdoors, and of using biomass as raw material, it could severely damage the environment. Others have analyzed the likelihood of an unstable arms race, and many have suggested economic upheaval resulting from the widespread use of free manufacturing. Some have even suggested that the entire basis of the economy would change, and money would become obsolete.

Sufficiently powerful products would allow malevolent people, either hostile governments or angry individuals, to wreak havoc. Destructive nanomachines could do immense damage to unprotected people and objects. If the wrong people gained the ability to manufacture any desired product, they could rule the world, or cause massive destruction in the attempt. Certain products, such as vast surveillance networks, powerful aerospace weapons, and microscopic antipersonnel devices, provide special cause for concern. Grey goo is relevant here as well: an effective means of sabotage would be to release a hard-to-detect robot that continued to manufacture copies of itself by destroying its surroundings.

Clearly, the unrestricted availability of advanced nanotechnology poses grave risks, which may well outweigh the benefits of clean, cheap, convenient, self-contained manufacturing. As analyzed in Forward to the Future: Nanotechnology and Regulatory Policy, some restriction is likely to be necessary. However, as was also pointed out in that study, an excess of restriction will enable the same problems by increasing the incentive for covert development of advanced nanotechnology. That paper considered regulation on a one-dimensional spectrum, from full relinquishment to complete lack of restriction. As will be shown below, a two-dimensional understanding of the problem—taking into account both control of nanotech manufacturing capability and control of its products—allows targeted restrictions to be applied, minimizing the most serious risks while preserving the potential benefits.