Continuing with CRN's series of five installments analyzing nanotechnology and risk...
Earlier this week, Part 1 gave an overview of existing nanoscale technologies, and Part 2 covered the risks of nanoscale technology. Part 3 (yesterday) described molecular manufacturing, while Part 4 (today) assesses the risks of molecular manufacturing. Part 5 will be a conclusion with recommendations.
Part 4: Risks of Molecular Manufacturing
Lack of knowledge is probably the major contributor to molecular manufacturing risk. The overly skeptical and politicized attitude toward molecular manufacturing in the U.S. during the past decade has created a situation in which almost no information is available; in fact, a very informative set of studies was removed from the 21st Century Nanotechnology Research and Development Act at the last minute through the efforts of a special-interest group.
Due to the almost complete lack of studies, there are several major uncertainties about molecular manufacturing. When will it be developed? CRN believes it will likely be between five and twenty years from now, but other estimates vary widely. How hard will it be to develop? A recent study (PDF) funded by the NASA Institute for Advanced Concepts concluded that a self-replicating system might be substantially less complex than a Pentium IV chip. But the cost might be anywhere from $100 million to $10 billion -- though we can say with confidence that the cost will rapidly decrease. Who will develop it? The military in any of several nations might be motivated to develop it openly or secretly. Any of a number of companies might do a study to construct a roadmap and cost estimate, discover that development fits their budget and business plan, and launch a project.
Planning for molecular manufacturing also is lacking. It does not appear in any roadmaps or forecasts. Only a handful of organizations, including CRN and the Foresight Institute, have even begun asking questions about what would constitute appropriate policy for a radical increase in manufacturing capability, and no good answers are available yet. If it is developed in less than a decade, the most common response will probably be panic.
A self-contained general-purpose manufacturing technology would have significant military implications. So would sudden, multiple-order-of-magnitude improvements in computers, robotics, and avionics. If these are combined into a single breakthrough, nations without access to the technology would have reason to feel very threatened; preemptive strikes appear to be a possible response. Once the technology is developed and proliferated, arms races could evolve rapidly and lead to accidental or deliberate conflict. If the military aspects of molecular manufacturing become too significant, it may be restricted in a variety of ways, preventing most companies and potential customers from participating in its benefits. Threat or outbreak of war would create even more problems and uncertainties.
Civilian access to molecular manufacturing also raises several risk-related policy issues. It would not be difficult to construct weapons of mass destruction with the technology, once it has developed to the point of automated self-contained factories; this implies that it may be kept closely guarded. On the other hand, the potential benefits from extensive positive use argue strongly for the desirability and near-inevitability of widespread access.
The technology will rapidly become easier to develop; and as the saying goes, if it is outlawed, only outlaws will have it. A wide range of policy outcomes can be imagined, from near-complete lack of control (as we have today in computers, where anyone can write and distribute malicious programs), through regulation, to near-complete control even of enabling technologies -- which would require unprecedented levels of surveillance worldwide. Each of these scenarios comes with its own set of risks. Without some idea of which will be tried and which can be stable, detailed analysis would be premature. However, the probability of random policy hitting on a good solution appears to be low, so further study in this area is urgently needed.
In almost any circumstance, intellectual property will be a major commercial issue. Molecular manufacturing has the potential to replace or supplant most forms of manufacturing, as well as most extraction activity and several infrastructures. With a potential orders-of-magnitude gap between cost and value, the ability to extract licensing fees could be worth significant fractions of the gross world product. Conversely, there will be huge economic incentives to bypass such fees. The ongoing struggle between the entertainment industry and file-sharing activity is a mild preview.
Further complicating the issue will be humanitarian and environmental imperatives, as foreshadowed by the debate over affordable pharmaceuticals. There is little doubt that fully utilized molecular manufacturing could save tens of millions of lives per year, but it will be difficult to develop policy that can accomplish this while satisfying commercial imperatives. CRN has proposed that something like the current division between commercial and Open Source software be adopted, allowing lifesaving products to be designed and built without licensing fees by hobbyists and altruists while retaining the right to make money on branded and polished commercial products. However, this would be a very delicate balance, and a combination of legal and black-market backlash would punish failure.
It was once thought that molecular manufacturing systems would be uncomfortably close to biological systems, able to remain functional and find resources in the wild, building exponentially more copies of themselves until the biosphere was consumed. However, as understanding matured from concept to architecture in the early 1990's, it became clear that molecular manufacturing systems would not have nearly the required functionality to run wild, either during development or during use. It is also clear that there is no reason to build general-purpose manufacturing into products, and doing so would be quite inefficient.
This is equally true for weapons: a non-replicating weapon built by a separate and special-purpose factory would be more destructive than a weapon that had to lug around an entire manufacturing system. To design a small self-contained free-range self-replicator would be a very difficult and pointless task -- harder than designing a nanofactory. Since it does not appear impossible, and since hobbyists have designed pointless computer viruses, it cannot be ruled out in the long term. But we do not consider it a primary concern (PDF).
Nevertheless, public worries about gray goo may be a major factor in shaping molecular manufacturing policy. Mainstream fictional portrayals of gray goo-like developments, including Michael Crichton's Prey, will work to keep the idea current. The fact that it does not appear to be theoretically impossible will make it hard to completely calm fears that molecular manufacturing will lead to it. (At the same time, ongoing attempts to assert that all of molecular manufacturing is impossible will extend the current period of confusion about the technology and its implications.)
Confronted with these issues, it may be tempting for business or government interests to try to postpone or prevent the technology. However, this would probably increase the risks in the not-too-distant future. A growing body of research on enabling technologies, the rapid development of computers, and worldwide awareness of the possible benefits, will ensure development before long -- if not in the mainstream, then from an unpredictable source.
In summary, the largest uncertainties about molecular manufacturing today arise from a near-complete inability to make policy or even to understand policy tradeoffs. The largest risks come from the likely consequences of bad policy. War between the world powers, uncontrollable black markets, widespread economic upheaval, lethal levels of artificial scarcity, massively destructive individuals, and government oppression are all likely consequences of policy that is hasty, simplistic, heavy-handed, or inadequate.
The best way to mitigate these risks and consequences is to engage in rapid and comprehensive studies of molecular manufacturing. CRN has published a list of "Thirty Essential Studies", each with multiple sub-topics, which we assert is the minimum required to get a handle on the issues. We strongly encourage individual and cooperative efforts to investigate these and related questions.
Tune in tomorrow for Part 5: Conclusion and Recommendations.