In the present climate, the news that US military researchers are redoubling
their efforts to develop the first fully robotic fighter planes will dismay many doves
(see 鈥淩obot wars鈥�).
Indeed, even those of us who think the armed assault on the
Taliban is justified are likely to react with scepticism and unease. We will
remember those smart laser-guided missiles that went astray during the Gulf War
and rightly ask whether it makes sense to continue shifting ever more military
executive power from humans to machines.
The US is already using uncrewed planes to spy on enemy forces. It also has
self-guiding missiles that will find and lock onto a target if they know what to
look for and are pointed in the right general direction. But uncrewed fighter
planes are conceptually different. Crucially, they will have the power to choose
their own targets from scratch, an innovation that puts us slap bang in sci-fi
territory where machines decide what and who to shoot鈥攁nd machines get
blown away, not humans.
Grave concerns about reliability and accuracy are inevitable. In 40 years of
research into artificial intelligence, scientists have conspicuously failed to
make a machine with anything like the comprehensive visual intelligence of a
human being. Programmers talk enthusiastically about 鈥渧isual recognition
systems鈥� that cannot compete with a toddler. Even the most powerful computers
have problems telling basic objects apart.
Advertisement
More subtly, machines have no morality or free will. If their guidance system
fails or they make the wrong decision and massacre a village, they cannot be
tried for crimes against humanity. And it would take a very clever lawyer to
indict those who set them loose for war crimes in which they played no direct
part. Even if precision was guaranteed, these robots could ultimately be
profoundly destabilising. Governments will be readier to rush into action if
they think there鈥檚 little risk their troops will fly home in body bags. And
otherwise timid citizens may be more willing to back them.
Not all would though. Many will recoil at the very idea of machines devoid of
conscience deciding what and who gets destroyed. And thanks to 24-hour rolling
TV news, the public now has a ringside view of every major conflict. They may
yet decide that letting robots do the killing is not just too risky but in the
end a little cowardly too.
