Brad Smith, Microsoft’s president, last week told the Guardian that tech companies should stop behaving as though everything that is not illegal is acceptable. Mr Smith made a good argument that technology may be considered morally neutral but technologists can’t be. He is correct that software engineers ought to take much more seriously the moral consequences of their work.
This argument operates on two levels: conscious and unconscious. It is easy to see the ethical issue in Microsoft’s sale of facial recognition technology to US Immigration and Customs Enforcement while the Trump administration was separating children from parents at the US’s southern border. The moral stance of more than 3,000 Google employees who protested about its Maven contract – where machine learning was to be used for military purposes, starting with drone imaging – with the US Department of Defense should be applauded. Google let the contract lapse. But people with different ethical viewpoints can take different views. In the case of the Maven contract, a rival with fewer qualms picked up the work. Much is contingent on public attitudes. Opinion polls show that Americans are not in favour of developing artificial intelligence technology for warfare, but this changes as soon as the country’s adversaries start to develop them. There is an economic aspect to be considered too. Shoshana Zuboff’s insight, that the exploitation of behavioural predictions covertly derived from the surveillance of users is capitalism’s latest stage, is key. What is our moral state when AI researchers are paid $1m a year but the people who label and classify the input data are paid $1.47 an hour.
However, the most difficult human skills to replicate are the unconscious ones, the product of millennia of evolution. In AI this is known as Moravec’s paradox. “We are all prodigious Olympians in perceptual and motor areas, so good that we make the difficult look easy,” wrote the futurist Hans Moravec. It is these that our brains excel in, hidden but complex processes that machine learning attempts to replicate. This presents the maker of such technology with a unique problem of accountability. If a building falls, the authorities can investigate, spot a failure, and put it down to engineering. We absorb the lessons and hope to learn from our mistakes.
But is this true for programmers? It is in the nature of AI that makers do not, and often cannot, predict what their creations do. We know how to make machines learn. But programmers do not understand completely the knowledge that intelligent computing acquires. If we did, we wouldn’t need computers to learn to learn. We’d know what they did and program the machines directly ourselves. They can recognise a face, a voice, be trained to judge people’s motivations or beat a computer game. But we cannot say exactly how. This is the genius and madness behind the technology. The promise of AI is that it will imbue machines with the ability to spot patterns from data, and make decisions faster and better than humans do. What happens if they make worse decisions faster? Governments need to pause and take stock of the societal repercussions of allowing machines over a few decades to replicate human skills that have been evolving for millions of years. But individuals and companies must take responsibility too.
Credit: Google News