| |
|
Lifeboat Foundation Nanoshield
Tomorrow's biggest danger may be nanoweapons (grey goo and other) created with molecular manufacturing. The Lifeboat Foundation proposes development of detection methods, such as infrared satellite surveillance for nanobot signatures, along with a three-layer defense system, with devices such as an orbiting mirror to focus concentrated sunlight on an ecophagic outbreak.
In Memory of Cassie Freitas
Originally published by Lifeboat
Foundation July 31, 2006. Reprinted with permission on KurzweilAI.net
August 6, 2006.
With participation by Amara
D. Angelica, Philippe
Van Nedervelde, Mike
Treder and other Lifeboat Foundation Scientific Advisory Board
members.
1. OVERVIEW
The most immediate danger facing life on earth is probably that
posed by biological weapons and emergent disease. The Lifeboat Foundation
BioShield
proposal [1], described by Lemelson-MIT Prize
winner Ray Kurzweil and US Senate majority leader Bill Frist, is
our recommended response to this danger. The BioShield proposal
emphasizes the development of technologies to combat bioweapons,
such as biological viruses, by developing broad tools to prevent
their development and to destroy them.
However, tomorrow's biggest danger is nanoweapons, and we believe
it is now time to develop a solution to this problem. As Ray Kurzweil
said, "As the threshold for self-organizing nanotechnology
approaches, we will then need to invest specifically in the development
of defensive technologies in that area, including the creation of
a technological immune system."
There are two types of nanoweapons:
1) Self-replicating weapons (ecophages) that make copies of themselves;
their only means of attack may be to "eat" the enemy or
his resources as they self-replicate.
2) Nonreplicating nanoweapons, similar to the tools of war today,
that are manufactured in a factory and then used in battle.
The NanoShield proposal has been designed primarily to handle
self-replicating weapons, but it will also be an excellent first
line of defense against nonreplicating weapons. Nonreplicating weapons
are more difficult to defend against, since they don't need to spend
a lot of time and effort in replicating. They are also easier to
design, since they do not have to include instructions on how to
replicate.
2. BACKGROUND
One of the earliest-recognized and best-known dangers of molecular
nanotechnology is the risk that artificial nanotech replicators
[2]—capable of digesting biological materials
and functioning autonomously in the natural environment—could
quickly convert the entire global ecosystem into more copies of
themselves.
This is a scenario often referred to as "grey goo," but
more accurately termed "global ecophagy," a term coined
by Robert A. Freitas, Jr. [3]. Such replicators,
called "ecophages," would constitute a class of sophisticated
artificial life forms more lethal than any plague that has ever
existed on this planet. If they are ever built and released, ecophages
will need to be controlled by a sophisticated artificial immune
system more powerful than any immune system that has ever been existed
in natural biology.
The human immune system does not have to recognize dangerous invaders
the way a nanotech system would. Our immune system merely has to
recognize non-invaders and attack everything else. Another important
distinction is that biological immune cells—and the invaders
that they must combat—will both replicate at biological speeds
and energy levels.
In contrast, a contest of exponentially growing numbers of nanotech
devices would cook the biosphere
in waste heat [3], especially if a large number
of novel replicating nanodevices were released simultaneously, and
if a different type of defensive device were needed to stop each
of them.
The human immune system also benefits from co-evolution with its
assailants. Microbes are selected to not overwhelm it too quickly,
otherwise plagues would burn out and the microbes would also lose.
The greatest immunological advantage of a human immune system may
be the vast number of humans in which it is found. When microbes
overwhelm it in one human, that human's selfish genetic material
can afford to simply die, while living on in other humans. But we
have only one earth, so we cannot afford such sacrifices of a global
scale.
For these reasons, we have a more difficult task than that facing
nature if we are to defend ourselves against artificial replicators.
But we also have advantages not possessed by nature. Most important
among these are the power to use design, to thoroughly analyze nanomachines
we find, and to employ macroscale phenomena in our defense.
3. THE NANOSHIELD PROPOSAL
Our proposal for a NanoShield encompasses five specific recommendations,
as follows.
3.1 THREAT DETECTION
To begin thinking about this problem, it is first necessary to
determine the incidence of ecophagy that is likely to be detectable.
This will be primarily a function of the pervasiveness of our defenses,
and of the efficiency with which they can identify ecophagy, bearing
in mind that ecophages may be intentionally designed to resist identification.
If nanorobots were withdrawing primarily carbon from the environment
to build diamond, you could in principle search for the surplus
or "waste" atoms that they discharge. For example, if
an ecophage was consuming CHON-based organic material, and removing
mainly the C atoms for incorporation into its mostly hydrocarbon-based
replicas, it would presumably be discharging the unused H, O, and
N atoms into the local environment as waste products in some form.
But trying to detect ecophages by searching for the waste atoms
offers several challenges:
If they discharge as they feed, the discharge could be hidden
by designing the ecophage to release mostly "natural"
appearing effluents. For example, waste O, N and H atoms could
be released as atmosphere-like O2, N2, H2,
or H2O.
Unless there were a lot of ecophages concentrated in a small area,
the volume of such effluent discharges would be relatively small,
and any wind could rapidly disperse the effluents, even if they
could somehow be recognized as artificial.
Ecophages could package their wastes into compressed-gas or solid-matter
pellets and then drop them into the dirt. If they were covered
with a camouflage coating, these droppings would be undetectable.
Some ecophages might be made of non-hydrocarbon ceramics such
as boron nitride or silicon nitride, and thus would have a different
effluent signature than expected for diamondoid ecophages. Such
devices might not even need to consume biology during their relatively
slow Build (replicating) phase, but could consume rocks, etc.
instead, and then only consume biology during their relatively
fast Destroy (nonreplicating) phase [3].
If the ecophage made a good effort to camouflage its effluents
they probably could not be detected, so a different detection method
would need to be tried to find the ecophage.
Two possible techniques are spectrographic analysis and sonographic
detection. One could detect MM products spectroscopically,
based on the presence of particular types of chemical bonds. However,
this could be defeated by ecophage designers by coating them with
something that looks natural, like silica (sand) or an outer shell
of some magnesium-iron-silicate-etc. mineral that looks exactly
like ordinary dirt.
MM products could be detected sonographically
based on their existence in the environment as multiple clusters
of matter that resonate at the same set of frequencies. However,
this would have the problem that nanorobots and their parts have
very high resonant frequencies—gigaHertz or teraHertz—because
they are so small. Also, acoustic waves of these frequencies are
hugely attenuated in their passage through air, or even water, so
their useful range would be very short, on the order of microns.
And it would not be wise to assume that ecophages will cluster into
nice large macroscale "tuning forks" that would make them
easier to detect—an ecophage designer would probably not require
his ecophages to do any aggregation at all in order to replicate.
Another method is to examine chunks of material. If we assume a
possible size on the order of 10 cubic microns, detection of a potentially
dangerous nanodevice in a cubic meter of material would require
that 1017 chunks of that material be examined. To perform
such an examination for each of the 1015 cubic meters
within two meters of the earth's surface, using rod logic (nanomechanical
computation at the molecular level, as proposed by K. Eric Drexler
in Nanosystems), would require roughly the Earth's incident
solar energy over a 15 minute period for every computational operation
involved in characterizing a particle as a threat.
It would also require the impractical disassembly of every object
on Earth and some technique for utilizing the information acquired.
Because some defensive probes would be expected to fail in the ordinary
course of events, this technique would also fail to detect ecophagy
armed to defend themselves. Even a less thorough examination, testing
a random sample consisting of one in 1012 regions of
space, would be disruptive, intensive, and relatively easy to circumvent.
Another method is to use three-dimensional images of nanoparticles
that have been obtained with a microscope using newly developed
coherent X-ray diffraction instead of focusing them [12].
This would allow for non-destructive forensic analysis to see if
it is an MM product.
Trying to detect MM products tactilely, based on hardness, could
be frustrated by camouflage coatings and would require physical
contact, which generates a large-numbers problem.
So rather than computationally analyzing and characterizing random
chunks of matter to determine whether they are capable of self-replication,
a better solution is to continuously monitor the heat signature
of the entire global surface and possibly the subsurface [3].
If this is combined with sophisticated pattern recognition, developing
trouble will be detected reasonably quickly. Manual inspection nanorobots
would then be sent to only those regions identified by the thermal
pattern recognition software as scoring high on the "possible
trouble" index. The atmosphere and oceans will need to be monitored
as well.
Is it possible for an ecophage to mask its infrared signature and
thus elude detection? Mechanical and chemical activities develop
waste heat, so eventually this heat must appear somewhere in the
environment. One strategy an ecophage could employ to evade detection
would be to transfer the heat from its site of activity to some
distant site, where the heat could be dispersed more widely and
thus would be harder to distinguish from background levels.
For example, a Peltier-effect cooling system (electronic refrigeration)
could transfer heat from the ecophages through an underground wiring
network to a distant distributed thermal radiator system, possibly
diluting the thermal signature by 1000:1 or more. Fluid-driven heat
pipes or a complex of fractal diamond pipes (diamond is an excellent
heat conductor) might also be effective. Another ecophagic strategy
might be to acquire the biological feedstock at a given surface
site but not process it there, transporting it instead to a distant
location where the thermal signature of chemical processing could
be better disguised—for example, in a processing facility located
deep underground.
All such approaches can still be defeated if a high-resolution,
high-sensitivity global thermal map has been created, good baseline
statistics have been collected for many years for both surface and
subsurface temperatures, and they are closely and continuously monitored
using sophisticated pattern-recognition software.
For additional safety, some random sampling of materials could
be done, in addition to simply monitoring heat signatures. However,
note that only a thorough examination (including partial disassembly)
of found objects will suffice to determine whether they are products
of molecular manufacturing (MM) or not.
3.2 NONSPECIFIC IMMUNITY DEFENSES
Instrumentalities should be put in place that constitute a general,
nonspecific response to any perceived ecophagic threat. For example,
inspection nanorobots could be deployed to any area that is suspected
of having any sign of possible ecophagic activity.
If ecophages are detected, there could be a response from pre-positioned
stores of generic defensive nanorobots manufactured by a global network
of defensive nanofactory stations that have been put in place well
in advance of the outbreak of the threat. These first-line defensive
nanorobots will have generic abilities to disable ecophages—e.g.,
sensor blinding, spray painting to ruin energy-producing solar cells,
and perhaps some capability of mechanical disassembly or physical
crushing, electric shock, e-beam irradiation, and so forth. These
defenses will buy time for the specific immunity defenses to kick
in.
3.3 SPECIFIC IMMUNITY DEFENSES
A second set of instrumentalities that should be put in place comprise
a specific, targeted immunity response to a perceived ecophagic
threat. These defenses would be designed to attack the particular
ecophage in question. They could not be launched until the ecophage
was identified and its weaknesses determined. A regular program
of collecting and inspecting nanorobots found in the environment
via sampling from randomly selected locations would help to establish
a statistical baseline on extant nanorobot populations and would
also provide an early warning of new nanorobotic capabilities that
are being fielded.
The ability to detect and identify an object implies the ability,
if necessary, to selectively deliver energy into that object. Ultrasound
in the proper resonant frequency could deliver destructive amounts
of energy into pre-specified and molecularly precise objects.
Specific surface chemistries can be attacked via the relevant chemical
reactions. Chemical bonds can also be broken by electromagnetic
quanta
at the correct frequency. For example, an ecophage that internally
employed mechanosynthetic
tooltips possessing Ge-C bonds might be disabled by exposure
to 21 THz infrared radiation, the approximate resonant stretching
frequency of the Ge-C dimer bond.
Similarly, exposure to GHz microwaves might pump unplanned energy
into purely mechanical components of nanorobots, such as the logic
rods in mechanical nanocomputers operating at GHz frequencies, thus
permanently damaging them, if these moving components incorporate
any unbalanced electrical charge or conductive pathways in their
molecular structure.
A methodology similar to that used by the human immune system is
another option. Surfaces complementary to those of undesirable environmental
contaminants, including MM devices, can be created and used to selectively
bind MM devices and isolate them. Sensors, solar cells and other
key parts of an MM device can be targeted as well. Such specific
anti-MM defenses would be launched at a detected infection, although
it is, of course, unlikely that permission would be given to carpet
the entire Earth (and its atmosphere and oceans) with them.
While molecular manufacturing systems must fight entropy to build
molecularly precise systems, countermeasures can work with entropy.
In other words, on the molecular level, as on any other, once detection
has occurred, destruction is far easier than creation and takes
much less time. As a result of this, except for ecophagic populations
much greater than the populations of countermeasure devices, the
time taken by countermeasures to eliminate ecophagic infestations
will be dominated by search time. Search time should usually be
inversely proportional to the concentration of targets.
For this reason, an exponentially replicating ecophage population
can be stopped by a constant-sized population of anti-ecophages,
or more precisely, a constant concentration (per volume) of anti-ecophages.
This implies two things: (a) You don't need to respond to an ecophage
outbreak instantly; and (b) You don't need to get into an exponential
race.
As MM populations in the environment are monitored, any non-Brownian
diffusion or rapid increase in incidence should flag the attention
of authorities who might examine the relevant population data, schemata
of the threatening nanodevice, and simulations of the device's behavior.
If they are concerned, they should authorize the release of countermeasures
stockpiled by a large and diffuse planetary grid of nanofactories.
Countermeasures need not be self-replicating, and in fact should
not be, since that would require them to be complex and slow; and
in addition, would raise the possibility that they could be preempted
for use as ecophages.
Division of labor is generally efficient, and the production of
countermeasures by specialized productive systems is an example
of this. Although a grid of productive nanofactories must be created
and stocked with feedstock and energy ahead of time for countermeasure
production, specific countermeasures need not be created until replicators
pose an immediate danger, so long as the total productive capacity
available for countermeasure production is sufficiently great.
Countermeasures could take the form of small molecules, nanomachines,
or macroscale devices such as ultrasound generators, sorting devices,
or even bush robots with target specific branch tips. ("Bush
robots" will have an immovable base that repeatedly branches
in a fractal way into trillions of nanoscale fingers. [4])
Collectively, specific countermeasures can be seen as the equivalent
of specific immunity. It could be innocuous, automatic, and continuous.
But unlike specific immunity in biology, ecophage countermeasures
can be subject to higher level analysis and centralized control,
enabling their modification to correct any unintended damage.
3.4 EMERGENCY DEFENSES
A third set of instrumentalities should be put in place that constitute
broad-brush emergency responses to a larger ecophagic threat. We
cannot rule out the possibility of rare situations in which the
normal nonspecific defenses fail, and successful specific defenses
cannot be mobilized. Examples of such a dire emergency would be
the existence of ecophagic replicators too numerous for cleanup
or the recognition that an uncharacterized ecophage or one with
no known specific countermeasures is replicating unexpectedly rapidly.
In such cases, it would be helpful if the NanoShield included emergency
defenses that would be effective against a wide range of ecophagy
types. With even broader impact than non-specific immunity responses,
the use of emergency defenses would disrupt lives and economic and
ecological systems. But the mere existence of these relevant defenses,
prepared but unused, will not cause harm.
Many of the proposed emergency responses will themselves cause additional
damage during the process of halting the ecophagic outbreak, much
as a surgeon's scalpel damages the tissue through which it cuts
during an operation to remove a more life-threatening tumor. For
this reason, emergency responses should be considered a last resort
and should be activated only in the direst of circumstances. In
the aftermath, advanced molecular manufacturing and nanomedicine
should allow us to repair many forms of damage to biological organisms,
including individual human beings. Much, although perhaps not all,
of the natural global ecological infrastructure might be reconstituted
if proper genetic and statistical records have been maintained that
describe the location and design of every large object and organism.
The possible misuse of specific countermeasures or emergency defenses
is inevitably a serious concern, but one that should be almost as
manageable as the risk of misuse of nuclear weapons. We say only
"almost as manageable" because MM
seems to favor evasion more than detection. This makes infiltration
and infiltration-dependent hacking of the defensive systems themselves
easier than it is in the case of nuclear weapons.
Also, unlike nuclear defense monitoring systems, such as Geiger
counters, anti-ecophagy defense systems leave footprints when taking
in information from the outside world, in the form of the data stream
from their continual monitoring activities.
Some examples of emergency ecophagy defenses are:
a) Skysweepers. Air-filtering nanoscoop devices could filter the
whole Earth's atmosphere, thereby removing all aerovores, as first
described by Freitas [3].
b) Utility Fog. Massive utility
fog curtains capable of establishing filters for the separation
of the atmosphere into compartments, containing an ecophagy outbreak,
or enabling the rapid establishment of expanding sterile bubbles
(barriers within which organisms can be safe from any ecophagy that
they don't bring in with them). Curtains may defend their integrity
with multiple layers, with sensors that can recognize damage and
respond with films of relatively inert substances that cannot be
modified by known mechanochemical
reactions (at room temperature or in general).
c) Solar Shades. Large solar shields that can be used to block the
sunlight reaching Earth's surface, thus denying power or reducing
the power available to solar-powered replicators. Prompt disablement
or sequestration over a short time period of ecophages rendered
dormant or lethargic may allow most terrestrial plant life to survive
the prolonged darkness unharmed.
d) Localized Heating. Localized heating increases thermal motions
in the mechanosynthetic tools used by the ecophage to build new
molecular structures, causing onboard temperature-sensitive mechanosynthetic
reactions to become unreliable. This would lead to fatal errors
in the fabrication and assembly of daughter ecophages and most probably
the permanent poisoning of the onboard mechanosynthetic tools. Localized
heating may be an inevitable side effect of the use of other specific
countermeasures at high power but can also be obtained more directly
by relatively simple measures.
For example, an orbiting mirror could be used to focus concentrated
sunlight into a specific ecophagic outbreak region, with the duration
and intensity of localized temperature increases carefully controlled
to maximize damage to the ecophages and minimize damage to the environment.
Alternatively, an orbiting laser beam could be directed onto the
ecophagic outbreak site (heating the ecophages). Ideally any temperature
changes would be confined to the smallest possible area.
e) Electromagnetic Pulse (EMP). Nuclear weapon explosions are known
to create very sharp pulses of high-intensity electromagnetic radiation
that can destroy electronic equipment. EMP can also be generated
by non-nuclear systems. Ecophages with onboard nanoelectronic components,
including sensors, computers, electrical motors or generators, and
power conduits, would be seriously damaged and probably rendered
entirely inoperative, if exposed to EMP. Only ecophages with all-mechanical
inner workings or that are heavily shielded would be immune to EMP
damage. Of course, many microelectronic and macroelectonic devices
that are not "hardened" (shielded and otherwise protected
against radiation) will be similarly damaged and would have to be
rebuilt in the aftermath, although EMP generators could be deployed
against ecophagic outbreaks in limited areas, using directional
antennas to minimize damage to electronic devices. One important
benefit of this approach is that EMP could be used against ecophages
infesting populated areas, without causing significant biological
damage to living things.
f) Radiation. Finally, high-power emitters of fairly penetrating
radiation, possibly x-rays or electrons from thermionic
emitters, can be used to destroy all large molecularly structured
systems within a large volume. Radiation can be tuned to minimize
interaction with organic tissue, particularly with key tissues such
as the nervous system, but basically this proposal relies on nanomedical
systems that can be rapidly deployed to repair nanoscale damage
before it brings about larger scale and more complex forms of damage.
This proposal may work well with (b), enabling organisms to be sterilized
while they enter quarantined compartments. Other methods of sterilization
include the use of nanomachines to remove all molecules from an
organism's body that are not pre-characterized as "normal."
This proposal is fairly similar to a generalized version of what
human immune systems typically try to accomplish, e.g., the removal
of everything except an enumerated list of molecule types, so the
immune system might actually be enlisted to aid in the identification
of nanosystems that natural immune cells have no way to attack.
Biocompatible surfaces are likely to be well characterized in nanomedicine,
so such surfaces can probably be identified by cleanup nanomachines
unless the ecophages have masked surfaces to evade detection.
3.5 NEW MONITORING AGENCIES
Each government participating in the NanoShield should establish
and fund a new monitoring agency analogous to existing governmental
agencies that already monitor outbreaks of computer viruses—most
notably the US Department of Homeland Security's Computer Emergency
Readiness Team (US-CERT), the world's premier public-sector computer
security monitoring agency [5]. Other analogous
monitoring efforts include the Tsunami Warning System [6]
operated by NOAA and the US National Weather Service, and the Spaceguard
[7] telescopic monitoring effort, which continuously
searches the skies for evidence of an approaching asteroid capable
of impacting the Earth.
The proposed new nanotech monitoring agencies would be charged with
initiating the early studies and preliminary implementation of the
NanoShield. Each country's agency should coordinate with the other
agencies and when they are ready to establish active defenses outside
their own countries, they should establish a lead body to handle
this.
The ultimate objectives of these nanotech monitoring agencies, as
originally noted by Freitas [3], would be:
"Initiating a long-term research program designed to acquire
the knowledge and capability needed to counteract ecophagic replicators,
including scenario-building and threat analysis with numerical
simulations, measure/countermeasure analysis, theory and design
of global monitoring systems capable of fast detection and response,
IFF (Identification Friend or Foe) discrimination protocols, and
eventually the design of relevant nanorobotic systemic defensive
capabilities and infrastructure.
"A related long-term recommendation is to initiate a global
system of comprehensive in situ ecosphere surveillance,
potentially including possible nanorobot activity signatures (e.g.,
changes in greenhouse gas concentrations), multispectral surface
imaging to detect disguised signatures, and direct local nanorobot
census sampling on land, sea, and air, as warranted by the pace
of development of new MM capabilities."
This would lead to various practical early-stage monitoring activities
that could be implemented today, including most importantly [3]:
"Continuous comprehensive infrared surveillance of Earth's
surface by geostationary satellites, both to monitor the current
biomass inventory and to detect (and then investigate) any rapidly-developing
artificial hotspots. This could be an extension of current or
proposed Earth-monitoring systems (e.g., NASA's Earth Observing
System [8] and disease remote-sensing programs
[9]), originally intended to understand and
predict global warming, changes in land use, and so forth—initially
using non-nanoscale technologies. Other methods of detection are
feasible and further research is required to identify and properly
evaluate the full range of alternatives."
4. UNSTABLE ARMS RACE: NONREPLICATING NANOWEAPONS
Molecular manufacturing also raises the possibility of horrifically
effective nonreplicating nanoweapons. The difference in purpose
between a nanotech weapon and an ecophage is that an ecophage is
meant to destroy biological matter, while nanotech weapons can have
far greater diversity of purposes, including killing only specific
parties. Ecophages must devote significant resources to replication,
whereas nanoweapons can focus solely on destruction. This means
that active nanoweapons can be far more dangerous per gram than
ecophages, and can act much more rapidly because they need not waste
time replicating.
As an example, the smallest insect is about 200 microns. This creates
a plausible size estimate for a nanotech-built antipersonnel weapon
capable of seeking and injecting toxin into unprotected humans.
The human lethal dose of botulism
toxin is about 100 nanograms, or about 1/100 the volume of the weapon.
As many as 50 billion toxin-carrying devices—theoretically
enough to kill every human on earth—could be packed into a
single suitcase.
Guns of all sizes would be far more powerful, and their bullets
could be self-guided. Aerospace hardware would be far lighter and
higher performance. Built with minimal or no metal, it would be
much harder to spot on radar. Embedded computers would allow remote
activation of any weapon, and more compact power handling would
allow greatly improved robotics.
Other possible nanoweapons include:
- Arbitrarily large numbers of any robot.
- Deuterium filters for separating deuterium from seawater.
- Microscale isotopic separation of uranium.
- Massive utility
fog banks that simply contain all movement in a large region.
- Computer viruses that make other people's nanofactories build
bombs.
- Inhalable or skin-penetrating machines that travel to the nervous
system, allowing outside sources to take over inputs or outputs.
- Massive nanofactories that consume a substantial fraction of
earth's CO2.
These ideas barely scratch the surface of what's possible.
An important question is whether nanotech weapons'both replicating
and nonreplicating'would be stabilizing or destabilizing. Nuclear
weapons, for example, could perhaps be credited with preventing
major wars since their invention. However, nanotech weapons differ
from nuclear weapons. Nuclear stability stems from at least three
factors. The most obvious is the massive destructiveness of all-out
nuclear war.
All-out nanotech war is probably equivalent in the short term, but
nuclear weapons also have a high long-term cost of use (fallout,
contamination) that would be much lower with nanotech weapons. Nuclear
weapons cause indiscriminate destruction; nanotech weapons could
be targeted. Nuclear weapons require massive research effort and
industrial development, which can be tracked far more easily than
nanotech weapons development.
Finally, nanotech weapons can be developed much more rapidly due
to faster, cheaper prototyping. Greater uncertainty of the capabilities
of the adversary, less response time to an attack, and better targeted
destruction of an enemy's visible resources during an attack all
make nanotech arms races less stable. Also, unless nanotech is tightly
controlled, the number of nanotech nations in the world could be
much higher than the number of nuclear nations, increasing the chance
of a regional conflict blowing up.
Bottom line: all problems that could be caused by nanotech weapons
might not be solvable by the NanoShield alone, but having the NanoShield
in place would provide an excellent first line of defense. We welcome
suggestions from the public on how to improve the NanoShield so
it can better handle nonreplicating nanoweapons.
4.1 RISKS OF NANOSHIELD
The risk that the NanoShield would malfunction and accidentally
destroy property or life on this planet can be made as close to
zero as desired by increasing the reliability and redundancy of
control systems. The greater and true risk of NanoShield implementation
is that it might be purposely abused by people. For example, a NanoShield
in malevolent hands could be used to oppress individuals, groups,
or entire countries.
To minimize this risk, authority to activate the NanoShield should
be distributed among as many responsible but competing interests
as is practical, consistent with the need for potentially rapid
decision making by parties who have demonstrated by past practice
that they are ready and willing to take decisive action if the need
arises. One good solution might be to have the NanoShield controlled
by a coalition of democracies, perhaps NATO.
Less ideal would be to vest control of the NanoShield in the hands
of a single strong democracy such as the United States or Australia.
A more dangerous outcome may occur if all democracies ignore this
vital issue and allow, by default, a dictatorship such as China,
or a small private group, or even a lone individual, to control
the NanoShield. It is unlikely that the UN can effectively administer
the NanoShield, due to structural problems, including its inability
to make rapid decisions, the veto power of non-democratic nations
having permanent seats on the Security Council, and the large number
of dictatorships represented among the UN membership.
5. INFOSHIELD PROPOSAL
To further increase global security and to lessen the need for
the most dangerous elements of the NanoShield to be activated, we
recommend consideration of the Lifeboat Foundation InfoShield
Proposal [10], to be implemented alongside
the NanoShield.
The InfoShield would deal with the problem that a powerful nanoweapon
could be developed in secret that could wipe out life on the earth
before the NanoShield could deal with it.
The InfoShield system would consist of multiple parallel, globally
deployed, nanotechnology-based surveillance systems, such as "smart
dust" (micro- or nanosize networked sensors that could
covertly detect anything). In addition, "sousveillance"
systems could be used. These would enable the public to turn the
tables and monitor the government (and perhaps others) via tools
like smart-dust data feeds—a possible checks-and-balance system
for the coming nano age.
5.1. TRANSPARENCY VS. PRIVACY
Of course, a smart dust system could also be used to abridge long-held
Constitutionally-protected rights to privacy. Special enabling legislation
or even an Amendment to the US Constitution might be required to
implement smart dust in a manner that would pass Constitutional
muster at the US Supreme Court. But as noted by Neil Jacobstein,
Chairman of the Institute
for Molecular Manufacturing: "Nanotechnology-enabled transparency
and accountability will produce the worst form of government, except
for all those other forms that have been tried from time to time."
Is it OK for governments to monitor civilians with quintillions
of sensors, and for civilians to monitor their governments with
quintillions of sensors? Or is that irresponsible and dangerous?
The bigger question is: do you want the NanoShield to try to handle
every possible nanoattack (we are not guaranteeing that it can handle
all nanoweapon types, although we are pretty confident that it can
handle ecophagy attacks)? Or do you want to try to stop hostile
forces from unleashing the attacks in the first place—which
would require extensive surveillance? Each attack stopped is one
less attack that might have overwhelmed the NanoShield.
The Lifeboat Foundation suggests that the InfoShield proposal is
a good first step in trying to prevent the attacks from happening
in the first place. (Nothing is likely to prevent 100% of the attacks,
which is why you will also need a NanoShield.)
But even combining smart dust with the three-layer defensive system
proposed for NanoShield cannot provide an absolute guarantee of
safety against all possible nanotech threats, especially given the
power that personal nanofactories [11], which
could be acquired by individuals, including terrorists. But NanoShield
should provide an excellent first line of defense, and adding smart
dust would further strengthen it.
6. CONCLUSION
Any particular ecophagy or nonreplicating nanoweapon defense can
be circumvented, but the number of people proposing ecophagy defenses
is likely to exceed the number building ecophages by many orders
of magnitude. A mix of defenses should be deployed, preferably by
multiple agencies, to minimize the risk of infiltration.
Some of these defenses should be announced publicly in order to
allow a hacker community to try their strength against them, as
is common with modern computer security, while other defenses should
be kept secret to avoid their circumvention. In this case, the total
barrier of a multilayer defensive system like NanoShield should
be sufficient to prevent the effective malevolent use of self-replicating
nanosystems, and should provide an excellent first line of defense
against the threat of even more potent nanoweaponry.
However, it is not necessary to implement the entire NanoShield
plan to be reasonably protected against ecophagic attacks. Even
a partial implementation would greatly increase the odds that an
ecophagic or nonreplicating nanoweapon attack would leave some survivors
and would easily be able to handle the bioweapon and pandemic problems
that the BioShield proposal [1] was developed to handle.
The reason the NanoShield could handle bioweapons and pandemic
problems is that the NanoShield would be designed to handle a large
range of designs, from carbon-based to silicon-based to boron-based,
to ecophages with virtually no onboard intelligence, as well as
those with onboard sophisticated computers, all-mechanical inner
workings or designs that include electronic components, etc. In
contrast, bioweapons and pandemics would have a much smaller range
of designs and therefore be easier for the NanoShield to defeat.
NOTES AND REFERENCES
1. Lifeboat Foundation Scientific Advisory Board,
Lifeboat
Foundation BioShield, July 2006.
2. Robert A. Freitas Jr., Ralph C. Merkle, Kinematic
Self-Replicating Machines, Landes Bioscience, Georgetown,
TX, 2004.
3. Robert A. Freitas Jr., Some
Limits to Global Ecophagy by Biovorous Nanoreplicators, with Public
Policy Recommendations, Zyvex preprint, April 2000.
4. Hans Moravec and Jesse Easudes, Fractal
branching ultra-dexterous robots, NASA: Advanced Concepts
Research Projects, January 1999.
5. United
States Computer Emergency Readiness Team (US-CERT).
6. The
Tsunami Warning System.
7. The
Spaceguard Central Node. See also: Spacewatch
Project.
8. EOSDIS
Earth Observing System Data Information System.
9. B. Lobitz, L. Beck, A. Huq, B. Wood, G. Fuchs,
A.S.G. Faruque, R. Colwell, Climate
and infectious disease: Use of remote sensing for detection of Vibrio
cholerae by indirect measurement, Proc. Natl. Acad. Sci.,
97(2000):1438-1443.
10. Philippe Van Nedervelde, Lifeboat
Foundation InfoShield, July 2006.
11. Robert A. Freitas Jr., Economic
Impact of the Personal Nanofactory, Nanotechnology Perceptions:
A Review of Ultraprecision Engineering and Nanotechnology 2(May
2006):111-126.
12. Eric D. Isaacs, X-ray
nanovision, Nature Vol. 442 (July 6, 2006):35,
© Lifeboat Foundation 2006
| | |
| |
Mind·X Discussion About This Article:
 |
|
|
 |
Lifeboat Nanoshield response
|
|
 |
|
Long post...
My comments & responses to various sections of the Lifeboat Nanoshield Proposal
In 2. Background,
2nd paragraph, "The human immune system does not have to recognize dangerous invaders the way a nanotech system would. Our immune system merely has to recognize non-invaders and attack everything else. Another important distinction is that biological immune cells ' and the invaders that they must combat ' will both replicate at biological speeds and energy levels."
Could there be a signature or other trusted identifier that would allow all "good" or trusted things to be recognized as such?
3rd paragraph (& I may just be being ignorant here): "In contrast, a contest of exponentially growing numbers of nanotech devices, by contrast, would cook the biosphere in waste heat [3], especially if a large number of novel replicating nanodevices were released simultaneously, and if a different type of defensive device were needed to stop each of them."
If this is so, why don't we now cook the biosphere through the exponential replication of biological organisms, such as bacteria?
4th paragraph: " Microbes are selected to not overwhelm it too quickly, otherwise plagues would burn out and the microbes would also lose."
Don't plagues burn out?
In 3. The Nanoshield Proposal,
Section 3.1, 5th paragraph (2nd inset paragraph): "Unless there were a lot of ecophages concentrated in a small area, the volume of such effluent discharges would be relatively small, and any wind could rapidly disperse the effluents, even if they could somehow be recognized as artificial."
Does the waste heat burning up the biosphere negate the idea of small effluent discharges (if we can call waste heat effluent here), or are we talking about small-scale number of ecophages, non-exponential?
Next paragraph: "Ecophages could package their wastes into compressed-gas or solid-matter pellets and then drop them into the dirt. If they were covered with a camouflage coating, these droppings would be undetectable."
Again - I think exponentially increasing numbers of anything will produce noticeable waste...unless again, we're talking about a smaller number of these things.
Paragraph 15: "Manual inspection nanorobots would then be sent to only those regions identified by the thermal pattern recognition software as scoring high on the "possible trouble" index. The atmosphere and oceans will need to be monitored as well."
This makes a lot of sense. But are normal environmental heat effects minimal and stable enough that such monitoring wouldn't either turn up mostly false positives, or (if tuned the other direction) miss relevant thermal events?
Also, discussion in & around paragraph 15 - ideas on how to solve the issues at stake:
One reason to keep a profit motive economy may be to encourage crowds of researchers / programmers / entrepreneurs to come up with solutions. Solving a problem in massively parallel fashion - by having lots of people & organizations trying, of their own volition, to solve the problems for profit - will produce a plethora of ideas and products that a few people won't.
3.2 Non-Specific Immunity Defenses
I would think that international participation and cooperation would b essential here. That is addressed later in the document, though.
3.3 Specific Immunity Defenses
My comments here aren't tied to specific paragraphs.
Comment related to 1st page of 3.3:
Open source designers hacking and beating on designs, as well as coming up with their own would be important here. Perhaps servers, pizza and Coke could be provided as incentives... :-)
Comments related to 2nd page of 3.3:
1: How can we be certain that immuno-defensive "bots" don't overwhelm the biosphere with waste heat (as discussed earlier in the document)? An attack by otherwise useless nanobots could be designed to do just that. The attacker sends in hordes of extra-cheap nanodevices that are alien but otherwise harmless, and the defense system jumps into play, generating tons of waste heat and self-damaging the defender.
2: An early immune system might be widely distributed and seeded into the environment which might as a side effect (or as a stealth justification) be used to clean up environmental hazards, toxins, spills, trash, etc. (Also, perhaps as a means to rid our bodies of some of the same. - as substitutes for today's common inoculations, e.g.) Of course, we would need to be pretty sure we know what we're doing to a complex biosphere. A stupid and simple example: clean up by destroying mosquitoes and frogs lose an important food resource, and frog-eating cranes lose an important food source, etc.
3: Regarding wide distribution of a grid of defensive productive nanofactories. Free (or all) Personal Nanofactories could have emergency overrides that would allow them to be used for defensive measures, including building additional nanofactories specifically for building defensive bots.
This is a good reason to give away PNs very early on once they come into existence / production. They become dual-use: feed, clothe, educate & house people everywhere, and act as an instant defense when needed as they are already everyplace by the time we have an attack. Then we don't need to pay or make the effort to put the exclusively defensive grid in place.
It seems that they would need to be designed to self-digest or self-destruct if other efforts were made to reprogram, hack, or otherwise repurpose said devices.
In 3.4 Emergency Defenses
3.4 e) Using EMP to disable "bad" nanodevices and then rebuilding in the aftermath is an intensely nontrivial task, In a populated area, we are talking about billions of devices: computers, watches, cars, Tivos, etc. It would bring the local economy to a standstill, not to mention thousands of lifesaving devices, automobile controls, radios so people can hear how to handle the emergency, cell phones, you name it. Maybe not so high a cost to pay when the alternative might be extermination, admittedly, but still a very high price.
3.4 f) "This proposal is fairly similar to a generalized version of what human immune systems typically try to accomplish, e.g., the removal of everything except an enumerated list of molecule types, so the immune system might actually be enlisted to aid in the identification of nanosystems that natural immune cells have no way to attack. "
Is this absolutely true? I may be just ignorant here, but is it possible that our immune system removes only specific threats or threat-like circumstances and leaves everything else alone?
3.5 New Monitoring Agencies
If truly an International effort from the get-go (not some kind of "coalition of the willing") then trust and buy-in will be built in. I think it's necessary if it is to be truly effective, given the amount of damage relatively small states might be able to do when enabled by MM.
4. Unstable Arms Race: Nonreplicating Nanoweapons
General discussion-
As with MAD (Mutually Assured Destruction), the threat of a massive nanotech response (MAND) might serve to deter attacks by states.
Having greater uncertainty wrt the capabilities of the adversary might actually be a more stable situation due to the possibility of the aforementioned MAND. I hate the idea as policy, but it seems to have worked with nukes (so far).
As for tightly controlling nanotech - the cat is out of the bag. There are too many nations and organizations working on it now to have any effective control, IMO.
4.1 Risks of Nanoshield: "Less ideal would be to vest control of the NanoShield in the hands of a single strong democracy such as the United States or Australia."
I agree with this assessment. It might be nice to see an explanation that others (especially within the US) would buy.
5. Transparency vs. Privacy
Are we ready for ubiquitous surveillance by anyone? Excuse me for being gross for a moment, but is everyone willing to be seen whacking off, eating their boogers, looking at porn, playing with a spouse (or a mistress), doing their taxes, wiping their butts, being alone, staying "in the closet"? Or planning wars?
What kind of processing power will it take to monitor all these inputs?
How are you going to get lawmakers to let it be two-way?
The document says (if I may excerpt), "The bigger question is: do you want the NanoShield to try to handle every possible nanoattack ... Or do you want to try to stop hostile forces from unleashing the attacks in the first place..."
This reminds me of "Are you with us or against us?"
Or - do you let us listen in to your conversations or are you with the terrorists?
There may be other choices. I don't know what they are at the moment, but perhaps this is the moment to ask, before we accept universal surveillance. It is arguable whether ubiquitous sensors are inevitable.
What are some other choices? |
|
 |
|
 |
|
|
|
|
Re: Lifeboat Foundation Nanoshield
|
|
|
|
I have translated into Russian 'Lifeboat Foundation Nanoshield' http://www.scribd.com/doc/12113758/Nano-Shield and I have some thoughts about it:
1) The effective mean of defense against ecofagy would be to turn in advance all the matter on the Earth into nanorobots. Just as every human body is composed of living cells (although this does not preclude the emergence of cancer cells). The visible world would not change. All object will consist of nano-cells, which would have sufficient immune potential to resist almost any foreseeable ecofagy. (Except purely informational like computer viruses). Even in each leaving cell would be small nanobot, which would control it. Maybe the world already consists of nanobots.
2) The authors of the project suggest that ecofagic attack would consist of two phases - reproduction and destruction. However, creators of ecofagy, could make three phases - first phase would be a quiet distribution throughout the Earth's surface, under surfase, in the water and air. In this phase nanorobots will multiply in slow rate, and most importantly, sought to be removed from each other on the maximum distance. In this case, their concentration everywhere on the Earth as a result would be 1 unit on the cube meter (which makes them unrecognazible). And only after it they would start to proliferate intensely, simultaneously creating nanorobots soldiers who did not replicate, but attack the defensive system. In doing so, they first have to suppress protection systems, like AIDS. Or as a modern computer viruses switches off the antivirus. Creators of the future ecofagy must understand it. As the second phase of rapid growth begins everywhere on the surface of the Earth, then it would be impossible to apply the tools of destruction such as nuclear strikes or aimed rays, as this would mean the death of the planet in any case - and simply would not be in store enough bombs.
3) The authors overestimate the reliability of protection systems. Any system has a control center, which is a blank spot. The authors implicitly assume that any person with a certain probability can suddenly become terrorist willing to destroy the world (and although the probability is very small, a large number of people living on Earth make it meaningful). But because such a system will be managed by people, they may also want to destroy the world. Nanoshield could destroy the entire world after one erroneous command. (Even if the AI manages it, we cannot say a priori that the AI cannot go mad.) The authors believe that multiple overlapping of Nanoshield protection from hackers will make it 100 % safe, but no known computer system is 100 % safe ' but all major computer programs were broken by hackers, including Windows and IPod.
4) Nanoshield could develop something like autoimmunity reaction. The author's idea that it is possible to achieve 100 % reliability by increasing the number of control systems is very superficial, as well as the more complex is the system, the more difficult is to calculate all the variants of its behavior, and the more likely it will fail in the spirit of the chaos theory.
5) Each cubic meter of oceanic water contains 77 million living beings (on the northern Atlantic, as the book 'Zoology of Invertebrates' tells). Hostile ecofages can easily camouflage under natural living beings, and vice versa; the ability of natural living beings to reproduce, move and emit heat will significantly hamper detection of ecofages, creating high level of false alarms. Moreover, ecofages may at some stage in their development be fully biological creatures, where all blueprints of nanorobot will be recorded in DNA, and thus be almost no distinguishable from the normal cell.
6) There are significant differences between ecofages and computer viruses. The latter exist in the artificial environment that is relatively easy to control - for example, turn off the power, get random access to memory, boot from other media, antivirus could be instantaneous delivered to any computer. Nevertheless, a significant portion of computers were infected with a virus, but many users are resigned to the presence of a number of malware on their machines, if it does not slow down much their work.
7) Compare: Stanislaw Lem wrote a story 'Darkness and mold' with main plot about ecofages.
8) The problem of Nanoshield must be analyzed dynamically in time - namely, the technical perfection of Nanoshield should precede technical perfection of nanoreplikators in any given moment. From this perspective, the whole concept seems very vulnerable, because to create an effective global Nanoshield require many years of development of nanotechnology - the development of constructive, and political development - while creating primitive ecofages capable, however, completely destroy the biosphere, is required much less effort. Example: Creating global missile defense system (ABM ' still not exist) is much more complex technologically and politically, than the creation of intercontinental nuclear missiles.
9) You should be aware that in the future will not be the principal difference between computer viruses and biological viruses and nanorobots - all them are information, in case of availability of any 'fabs' which can transfer information from one carrier to another. Living cells could construct nanorobots, and vice versa; spreading over computer networks, computer viruses can capture bioprinters or nanofabs and force them to perform dangerous bioorganizms or nanorobots (or even malware could be integrated into existing computer programs, nanorobots or DNA of artificial organisms). These nanorobots can then connect to computer networks (including the network which control Nanoshield) and send their code in electronic form. In addition to these three forms of the virus: nanotechnology, biotechnology and computer, are possible other forms, for example, cogno - that is transforming the virus in some set of ideas in the human brain which push the man to re-write computer viruses and nanobots. Idea of 'hacking' is now such a meme.
10) It must be noted that in the future artificial intelligence will be much more accessible, and thus the viruses would be much more intelligent than today's computer viruses, also applies to nanorobots: they will have a certain understanding of reality, and the ability to quickly rebuild itself, even to invent its innovative design and adapt to new environments. Essential question of ecofagy is whether individual nanorobots are independent of each other, as the bacteria cells, or they will act as a unified army with a single command and communication systems. In the latter case, it is possible to intercept the management of hostile army ecofages.
11) All that is suitable to combat ecofagy, is suitable as a defensive (and possibly offensive) weapons in nanowar.
12) Nanoshield is possible only as global organization. If there is part of the Earth which is not covered by it, Nanoshield will be useless (because there nanorobots will multiply in such quantities that it would be impossible to confront them). It is an effective weapon against people and organizations. So, it should occur only after full and final political unification of the globe. The latter may result from either World War for the unification of the planet, either by merging of humanity in the face of terrible catastrophes, such as flash of ecofagy. In any case, the appearance of Nanoshield must be preceded by some accident, which means a great chance of loss of humanity.
13) Discovery of 'cold fusion' or other non-conventional energy sources will make possible much more rapid spread of ecofagy, as they will be able to live in the bowels of the earth and would not require solar energy.
14) It is wrong to consider separately self-replicating and non-replitcating nanoweapons. Some kinds of ecofagy can produce nano-soldiers attacking and killing all life. (This ecofagy can become a global tool of blackmail.) It has been said that to destroy all people on the Earth can be enough a few kilograms of nano-soldiers. Some kinds of ecofagy in early phase could dispersed throughout the world, very slowly and quietly multiply and move, and then produce a number of nano-soldiers and attack humans and defensive systems, and then begin to multiply intensively in all areas of the globe. But man, stuffed with nano-medicine, can resist attack of nanosoldier as well as medical nanorobots will be able to neutralize any poisons and tears arteries. In this small nanorobot must attack primarily informational, rather than from a large selection of energy.
15) Did the information transparency mean that everyone can access code of dangerous computer virus, or description of nanorobot-ecofage? A world where viruses and knowledge of mass destruction could be instantly disseminated through the tools of information transparency is hardly possible to be secure. We need to control not only nanorobots, but primarily persons or other entities which may run ecofagy. The smaller is the number of these people (for example, scientists-nanotechnologist), the easier would be to control them. On the contrary, the diffusion of knowledge among billions of people will make inevitable emergence of nano-hackers.
16) The allegation that the number of creators of defense against ecofagy will exceed the number of creators of ecofagy in many orders of magnitude, seems doubtful, if we consider an example of computer viruses. Here we see that, conversely, the number of virus writers in the many orders of magnitude exceeds the number of firms and projects on anti-virus protection, and moreover, the majority of anti-virus systems cannot work together as they stops each other. Terrorists may be masked by people opposing ecofagy and try to deploy their own system for combat ecofagy, which will contain a tab that allows it to suddenly be reprogrammed for the hostile goal.
17) The text implicitly suggests that Nanoshield precedes to the invention of self improving AI of superhuman level. However, from other prognosis we know that this event is very likely, and most likely to occur simultaneously with the flourishing of advanced nanotechnology. Thus, it is not clear in what timeframe the project Nanoshield exist. The developed artificial intelligence will be able to create a better Nanoshield and Infoshield, and means to overcome any human shields.
18) We should be aware of equivalence of nanorobots and nanofabrics - first can create second, and vice versa. This erases the border between the replicating and non-replicating nanomachines, because a device not initially intended to replicate itself can construct somehow nanorobot or to reprogram itself into capable for replication nanorobot.
|
|
|
|
|
|
|
