Maximise human safety vs Keep technology under control
Hot-headed, irrational and swayed by emotion – who’d want a human in control? If we could build machines capable of making tough choices for us, surely we should. That’s the line taken by people like roboticist Ron Arkin at the Georgia Institute of Technology in Atlanta. For Arkin, autonomous weapons – or killer robots – that remain rational under fire and behave exactly as they were trained to would be more humane than human soldiers in a war situation, and would save lives. We therefore have a moral imperative to create them.
Read more: The ethics issue – The 10 biggest moral dilemmas in science
Science has given us the power to design life, reshape the planet and colonise other worlds. But should we? 91av grapples with the big ethical questions
The same reasoning can be applied to many scenarios where human nature may stop us doing the right thing, from driving to making life-or-death decisions in hospitals to criminal sentencing. Computers are already moving into all these areas, and in many cases surpass humans where it counts. But how much autonomy should we give them?
Advertisement
The problem with fully autonomous machines from a moral point of view is that they cannot take responsibility for their actions. Human ethics is built on the assumption that actions are done by agents with the capacity to make a call between right and wrong. If we offload those actions on to machines, who do we blame when something goes wrong?
Filippo Santoni de Sio, a philosopher and ethicist at the Delft University of Technology in the Netherlands, calls this the “responsibility gap”. For him, it is crucial that a human always takes responsibility. This might be the machine’s designers or a designated human handler, much as a parent takes responsibility for their child or an owner for their pit bull. “It is morally wrong to give autonomy to machines because they cannot perform moral thinking,” says Santoni de Sio.
Arkin and others have suggested that at some point in the distant future machines capable of moral reasoning could be built. But even in such a case, many will still feel it is wrong to delegate such decisions to a machine. With people, we don’t just want them to do the right thing; we also have a need to understand why they act the way they do.
Blame is as much about explanation as it is compensation or retribution, says Santoni de Sio. “This is the main idea behind moral responsibility – we want to blame people not necessarily because we want to punish them, but because we want to understand. We want to see the person’s face, we want to ask questions.”
So where do we go from here? Where the stakes are highest, many are calling for an outright ban. Campaigners are pushing the UN to draw up a treaty curbing development of autonomous weapons, for example. Legislation is on its way for autonomous vehicles too. Germany has proposed a highway code for driverless cars, laying down guidelines for who is ultimately responsible in an accident. And similar regulations will be drawn up for medical robots.
Attitudes may change as we work through the moral maze surrounding autonomous machines. But for now, at least, we will be keeping them on very short reins.
Now that you’ve read the article, let us know what you think about this topic. Where do you stand?
This article appeared in print under the headline “Should we… Give robots the right to kill?”
