The U.S. military’s drones are defenseless against enemy fighters — as an incident this month over the Persian Gulf shows. But that wasn’t always the case. In 2002 the Air Force fitted some of its early-model Predator drones with short-range Stinger air-to-air missiles. But even with the right weaponry the robots were likely a poor match for enemy fighters, and the Air Force ultimately stripped them of the missiles.

That decision had ramifications a decade later on Nov. 1, when a pair of Iranian Su-25 attack jet pilots tried their damnedest to shoot down an unarmed Predator conducting what Pentagon spokesman George Little dubbed “routine surveillance” over the Persian Gulf 16 miles off the Iranian coast.

The Predator fled for safer airspace, but not before the Iranians made two passes firing 30-millimeter cannons.

It’s not clear that the now-removed defensive armament would have made any difference in last week’s jet-on-drone violence. The one instance of an air-to-air-armed Predator fighting back against an attacking aircraft resulted in the drone’s destruction. “Best to say that results were inconclusive,” retired Air Force general Dave Deptula tells Danger Room.

The results may have proved uninspiring, but the addition of Stingers to the lightweight Predator was an impressive bureaucratic feat. As journalist Richard Whittle described in a recent study, in 2002 the Air Force flew Predators over southern Iraq in support of the pre-war no-fly zone. Intervention by Iraqi MiG fighters made it hard for the drone crews to complete their surveillance missions.

A highly-regarded Air Force technical troubleshooter named James Clark approached Gen. John Jumper, chief of staff of the flying branch, with the idea of adding the 34-pound, heat-seeking Stingers to the drones.

Expectations were modest. “A Predator crew would find it hard, if not impossible, to spot an Iraqi fighter and launch a Stinger quickly enough to have a chance of hitting it, given the speed of a jet,” Whittle wrote. “At the very least, though, giving the Predator a way to shoot back might spook [Iraqi dictator Saddam] Hussein’s pilots, Jumper and Clark agreed.”

Predator-maker General Atomics got a contract to integrate Stingers in September, and by October a drone was flying mock dogfights against a Cessna playing the role of an enemy plane. Four Stinger test firings resulted in two misses.

In November the Stinger-armed Predators deployed to the Middle East and were soon maneuvering against Iraqi MiGs, albeit apparently without launching missiles. Then on Dec. 23, 2002, an Iraqi MiG-25 intercepted one of the Predators and, for the first time, the two planes exchanged fire.

Video of the brief battle, shot by the Predator, was obtained by CBS News. “The engagement began when the MiG turned to attack head-on and fired a missile,” Whittle recalled. “The Predator crew fired a Stinger back. The video shows the smoke trails of the missiles crossing, then the Stinger starting to dive, coming nowhere close to the MiG. Then the Predator video suddenly ends.”

The drone fell to the ground in pieces. The MiG was reportedly unscathed.

When U.S. forces invaded Iraq just over three months later, Hussein ordered the MiG-25s and the rest of his air force buried in the sand rather than face American F-15s. Predators continued flying over Iraq and Afghanistan. “Given the absence of air threats in Iraq and Afghanistan, the decision was made not to equip those RPA with air-to-air weapons,” Deptula says.

As a consequence later Predator models and their larger Reaper cousins would be unable to defend themselves against future foes not so eager to bury their air forces. But it’s not clear that drone air-to-air armament is worth it even now. “The predominant use of RPAs [Remotely Piloted Aircraft] over the past decade has been passive [intelligence] collection coupled with air-to-ground strikes in permissive airspace,” Air Force spokesman Lt. Col. Tadd Sholtis tells Danger Room. “There’s very little about their current capabilities — from their speed to their maneuverability to the range of visibility afforded to operators — that encourages operators to arm or employ them for air-to-air engagements in defended airspace.”

Peter Goon, an Australian aerospace analyst and drone skeptic, dismisses the 2002 dogfight as a deceptively simple “canned scenario” with unusually few variables. If anything, Predators and other RPAs were and are even less effective against enemy fighters than the Iraq battle implies.

True robotic air-to-air capability requires high-performing drones with ultra-sophisticated artificial intelligence, Goon argues. That could be “decades away, if [it happens] at all,” he tells Danger Room.

The military agrees. The Navy and Air Force have both begun thinking about so-called “sixth-generation” fighters that the services say could be unmanned. But these potentially robotic warplanes won’t enter service before 2030, according to current plans.

Until then, drones will be at the mercy of manned fighters. And adding missiles might not help.