It is remarkable to notice how often very intelligent people say the daftest things.
Over a thousand high-profile artificial intelligence experts and eminent scientists – including Professor Stephen Hawking – have signed an open letter warning of a “military artificial intelligence arms race” and calling for a ban on “offensive autonomous weapons.”
Actually, all weapons are offensive: a radar shield, for example, might be considered to provide such a good defence that it encourages the defender to go on to the attack. But leave that aside for a moment.
Their letter says: “Technology has reached a point where the deployment of autonomous weapons is – practically if not legally – feasible within years, not decades, and the stakes are high: autonomous weapons have been described as the third revolution in warfare, after gunpowder and nuclear arms.”
The authors argue that AI can be used to make the battlefield a safer place for military personnel, but that offensive weapons operating on their own would lower the threshold for going to war and result in greater loss of human life. Should one military power start developing systems capable of selecting targets and operating autonomously without direct human control, it would start an arms race similar to the one for the atom bomb. Unlike nuclear weapons, however, AI requires no specific hard-to-create materials and will be difficult to monitor.
In philosophical ethics, there is a famous rule: “Ought implies can.” In other words, you can only be obliged to do those things which you are capable of doing. The writers of that letter say there ought to be a ban on AI weaponry. The first question, therefore, is of who is to institute and police such a ban?. Let us say the United Nations, always supposing all members of the security council agreed to it. The next question is why should any nation state accept the ban? Nation states act in their own perceived interests and so, if the leaders of a particular country assessed that AI weapons would give an advantage over potential enemies, they would naturally proceed to manufacture AI weapons.
They could not afford not to. No responsible government can allow advantage to the enemy. And, should a government permit such an advantage – thus endangering the lives of its people – it would justly earn the people’s condemnation.
The case of AI weapons is technologically new, but it is not ethically new. It has happened time and again with the development of armaments from the crossbow to gunpowder, from the tank to the hydrogen bomb.
I suggest that the eminences who signed that letter confuse the possession of new weapons systems with their use. For again, a nation would not deploy a particular weapon if to do so would not be in its own best interests. For example, the great powers possess thousands of nuclear weapons, but only two atomic bombs have ever been used in warfare over the last seventy years. Why not? Not because some fantasists in CND have managed to ban them, but because to deploy them would be to invite destruction.
I know all this is not nice. It is not something which appeals to idealists. But idealism is not appropriate in a world that is far from being ideal.
Statesmen have a duty to deal with the rough-hewn world as it is, with all its messiness, compromises and blurred edges.
It’s called making the best of it.