The US military relied on its Predator and Global Hawk unmanned aerial vehicles (UAVs) for reconnaissance and identifying targets for the Afghan campaign.

"We're really developing some incredible technological ability on these unmanned vehicles," Beal says. "The next step is weaponising them."

In fact, that has already happened - Predators were carrying armour-piercing Hellfire missiles in Afghanistan.

"It's the first use of an unmanned aerial vehicle carrying air-to-surface missiles, ever," Beal says.

Unmanned aircraft have appeal: they're cheaper than manned warplanes and countries don't risk losing pilots on dangerous missions.

While the technology is new, the concept of recruiting automated machines for combat is an old idea.

Just before the end of the World War II the Germans were developing an explosive packed remote-controlled bomber airplane to deliver a devastating blow to British targets.

Twenty years later, in Vietnam, the US military strapped a bomb onto a UAV called the Firebee in an experiment with limited success.

In 1973, Israel began dabbling in remote-control 'drones' to confuse enemy radar. By the 1980s these UAVs were used in southern Lebanon and along the Israel-Syria border.

Today some 50 US companies, research institutions and government agencies are developing more than 150 UAV designs.

The US is the world leader in these automatic systems, distantly followed by Israel, Britain, France and Germany, and to a much lesser extent Russia and China.

It may be only a matter of time when man's role is diminished.

"The weakest link in the chain will be the man-in-the loop," Beal says. "There will be greater and greater pressure as machines become more sophisticated to hand over decision-making to machines because they will be able to process information much quicker than humans."

This too is already happening.

Israel's latest piece of military hardware, the Harpy - a 'fire and forget' assault UAV that detects, attacks and destroys radar emitters and missile sites - is already sold abroad.

The US Air Force and Lockheed Martin are developing a flying robot called locaas (Low-Cost Autonomous Attack System), a metre-long drone that can analyse targets and choose different warheads depending on target type. The military is even studying life forms to improve LOCAAS attack capabilities - including mimicking insects to make robotic drones 'swarm' their target.

And devices the size of pigeons are being developed in the USA by the military for surveillance and combat.

The leader in 'robotic warfare' is the US Navy. Within 15 years, half of the warplanes aboard US aircraft carriers will be robotic warplanes, experts say.

Leading the charge is Pegasus - the name of the robot aircraft being developed by Northrop Grumman under contract from the US Navy and Defence Advanced Research Projects Agency.

The kite-shaped automatic Stealth aircraft is designed for surveillance, air defence suppression and strike missions. Unlike its remotely piloted predecessors a human won't guide Pegasus.

Northrop Grumman aims to have its Pegasus prototype flying by 2002. The US Congress and US Navy want these unmanned combat aircraft from American aircraft carriers by 2010 to 2015.

Robots are also being developed for the oceans and ground - although such devices aren't as advanced as their airborne cousins. Naval research is exploring minesweeping robots that can detect and detonate submerged explosives. The US Marine Corps are also developing ground robots for urban warfare.

Using systems that act on their own has its downsides.

Beal says this technology "will change the nature of warfare in a very real way" and raise legal and military issues about the role of robots in combat.

"The problem comes when you really start giving command authority to machines that can make lethal decisions," he says.

But one robot expert doesn't think it is too much of an ethical quandary.

Anne Foerst, a professor of computer science and theology at New York's St Bonaventure University, says it's a question of who is responsible for a robot's actions.

"When robots are used on the battlefield and kill people they are not responsible - it's the humans that use them," Foerst, says, who formerly worked at the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology, a leader in developing robot systems.

Forest recognises the military's contributions in advancing robot technology - such progress can transfer to non-military applications that help humankind.

Her problems lie with how the military uses robots.

"For me, a robot carrying a weapon is big trouble," she says. "Our sense of who is the enemy is a social construct... and it is hard to translate for a robot."

Those 'social constructs' include appearances, attitudes, beliefs and behaviours.

"To find out what to shoot at, a robot would need as complex an understanding of social interactions and social groups as we do, because all that is needed to determine who is the enemy," Foerst said.

She says it will be a long, long time before robots can make such decisions.

Rear Admiral Richard Cobbold, director of Britain's Royal United Services Institute for Defence Studies, agrees.

"The way war is going, the human being is still going to be required because the nature of conflict is such that human contract - particularly at lower levels - is going to remain a key component," he said.

"Robots are not going to take over the world for a very long time," Cobbold says, "but the reliability of robots and artificial intelligence will complement the contributions of human judgement in the future." —Gemini News

Photo caption — UC Berkeley professor Dr. Ron Fearing holds a lifesize plastic mockup of the piezoelectric-powered robotic fly now undergoing development by a team of researchers. The tiny robo-insect will fly for 10-20 minutes at a time on wings that mimic those of a blowfly.