BioEd Online

At the 24th Army Science Conference, held in Orlando, Florida last December, Ray Kurzweil gave a keynote address entitled "Warfighting in the 21st Century". News@nature quizzed this renowned commentator on robotics about his views on future warfare.

How will warfare change in the next 50 years?

Within 25 years, non-biological intelligence will match human intelligence in areas in which humans now excel, principally in pattern recognition. It will combine these abilities with the inherent advantages of machine intelligence, such as speed, easy sharing of knowledge and skills. One implication will be that we will enhance our own biological intelligence through direct connection with non-biological intelligence.

Already, our abilities benefit from close collaboration with machines. Within 50 years, the non-biological portion of the intelligence of our civilization will predominate. Applying non-biological intelligence to areas such as strategy, decision-making and intelligent weapons will characterize military power.

You imply that one of the primary aims will, or should, be to remove humans from direct combat. How would that happen?

The US army's Armed Predator, a small, unmanned plane equipped with a laser, is a harbinger of an important trend. There is no reason why a human has to be inside a weapon. He or she can be in an appropriate virtual-reality environment and control the weapon from there.

This allows the weapon to be smaller, to take much greater risks and to dispense with life-support systems. It also puts the onus on secure communications. These weapons will develop increasingly sophisticated abilities to make their own tactical decisions.

Where will the future battlefields be? Will they include cyberspace?

One major development will be swarms of nano-engineered devices. Already, the US Department of Defense's Smart Dust project has prototypes that can reconnoitre. For example, millions of sensors can detect detailed movement of humans and machines in a large area. Ultimately they will be able to explore substantial terrains stealthily, perhaps looking for specific individuals. In a decade, they will be capable of performing offensive missions.

Cyberspace will be a key battlefield. With software running in and around our bodies and brains, software integrity and security will a key strategic issue.

How does your vision of robotic, virtual-reality warfare square with the notion that many modern wars involve terrorism or suicide attacks in civilian settings?

Centralized structures such as buildings, aeroplanes and cities are targets for terrorism, but decentralized, self-organizing systems are far more secure: meetings conducted through the Internet will not be subject to physical attack, and these will move towards full immersion virtual reality environments. And if we devise decentralized sources of energy, for example, nanotech solar panels and nanoengineered fuel cells, these will be less easy to disrupt. Decentralization of our infrastructure will result in suicide bombers becoming less of an issue. On the other hand, the ability of a bioterrorist to create a bioengineered virus will introduce new dangers. There are emerging strategies for dealing with this - for example applying RNA interference to combat newly introduced viral pathogens - but we need to increase the priority of developing these defences. Similar issues will arise when we have full molecular nanotechnology.

You suggest that one key concept is decentralization. Is there any reason to believe that military leaders will grant their 'agents' the necessary autonomy?

We are already seeing increased deployment of special forces that include intelligence agents who are empowered to make significant, on-the-ground decisions. Military tactics have largely moved beyond large massed armies.

The same applies to information and secrecy. Self-organization typically relies on local rather than global information, but will that require shifts in military thinking?

One key issue is giving tactical authority to machine intelligence. There is a hesitancy to do this, but as machine intelligence gains in sophistication, the line between strategic and tactical decision-making gradually changes. In 20 years, a human commander is likely to give her swarm of weapons a command to "Take that hill," leaving the tactical details to the swarm.

Do you have the impression that the military is genuinely ready to change the way it operates? There must be a lot of inertia in such a system.

The unexpected success of Armed Predator, and similar projects, has had a major impact. Technology typically moves in an incremental fashion rather than a single leap, but with each generation of technology taking less time than the last, this evolutionary process is moving more quickly.

Wars of conquest or permanent occupation seem increasingly unlikely. Granting that technology will change, what will wars themselves be about?

There was a major trend towards democracy in the 1990s, fuelled by decentralized electronic communication. We're not fully there but I believe this trend will continue. So a key struggle will be against fundamentalism, against groups animated by a desire to keep things the way they are, or were. These struggles will tend not to be classical struggles between nation states.

Where should a nation's priorities lie: in avoiding war, or in strengthening defences?

One can argue that these are not incompatible. Many of the elements of avoiding war are not military issues, but ones of promoting political stability, education, and other elements of societies in which all members have a stake. From a military angle, we will need to strengthen international cooperation, as modern dangers are global issues. Missiles, biological viruses, and information pathogens all travel around the world without regard for national boundaries.

What would you say to sceptics who argue that vast technological superiority has not prevented the US from suffering significant casualties in its current conflict?

The order of magnitude of US casualties in Desert Storm, the Balkans and the recent Iraq war are around a thousand deaths; Vietnam produced on the order of 100,000 US deaths, and casualties in earlier wars were much higher.

Perceptions of casualties are skewed because each event is brought to our living rooms in high resolution In the Second World War, there were single battles with tens of thousands of casualties that generated only grainy newsreels seen weeks later in movie theatres.

Does new technology, and in particular the 'singularity' that you have identified in its rate of change, make war more or less likely?

I think classical wars will occur less often. Decentralized communication such as the Internet is inherently a democratizing force, and has been behind the move towards greater democracy in the world. Although not a perfect system, democracies tend not to fight wars against each other. Future conflicts will tend to be with smaller groups who will try to amplify their impact by finding vulnerabilities in our technological infrastructure.