I’M SORRY, DAVE, I’M AFRAID I CAN’T DO THAT
From 2001’s Hal computer to Arnold Schwarzenegger’s Terminator and TV and film’s Westworld, it’s fair to say we’re uneasy about what robots might do to us when our backs are turned. The title to this piece represents the moment in 2001 when the computer turns against the astronaut to devastating effect. Our anxieties stem from the deepest nightmares about the creation of things that suddenly oppose, and overpower, us. In his analysis of future risk, Australian philosopher Toby Ord in his book The Precipice identifies unaligned artificial intelligence as one of the greatest threats to the future of the human race.
With the clear and present danger from pandemics and climate change, the Terminator risk still seems far fetched to many, but we would do well to focus our attention on today’s algorithms, and especially adaptive algorithms that learn on the job to make their own assessments. Dystopian futures may still be mocked, but the fumbling of UK A level results in 2020 showed what a badly designed algorithm is capable of and how it can mess up a generation at a stroke. It was a late wake up call; an outlier of what is happening under our noses.
Machine learning, or adaptive, algorithms can develop their own knowledge and can act in ways their inventors never intended and are increasingly gatekeepers for social goods. Job applications are frequently sifted by algorithm and it is estimated that 72% of CVs are ‘never seen by human eyes’. Machines are also used to judge a person’s character and ability by deep trawls of the personal data made available online – browsing history, views and connections made on social media. Latent weaknesses, like addictions, are illuminated.
The more we engage in our social media feeds, the more we are giving away to sources we didn’t even know exist. We are sanguine about the exchange of personal privacy online for a chance to make connections with others we could only dream of a few years ago, the assumption being that our data is only being used to sell us stuff we might actually like. The reality is different. Algorithms use every granule of our personal data to decide credit scores and access to tenancies. They may already be used to determine how much we pay in life insurance premiums or motor insurance. Ad this is only the beginning.
In his excellent ‘Future Politics’, barrister Jamie Susskind points out that machine learning algorithms can only learn from the data to which they are applied. He cites a competition in which six hundred thousand entrants globally sent selfies to be judged by an adaptive algorithm. Of the forty-four faces judged to be the most attractive, thirty-eight were white and only one had visibly dark skin. Susskind concludes:
No matter how smart an algorithm is, if it is fed a partial or misleading view of the world it will treat unjustly those who have been hidden from its view or presented in an unfair light.
Most coders will be well-intentioned, but are subject to unconscious – as well as conscious – biases that feed into their work. Our awe at the speed and efficiency of computers is matched by our credulity that they will give us the impartiality we long for in making public policy decisions.
There are no grounds for Luddism, for these developments are here to stay. We cannot un-invent the algorithm and it will, in any case, do much good. Instead, our ethical debate should catch up with technological developments. Right now, the car is speeding away and we are running breathlessly behind it.
Human justice is flawed, but flexible and capable of layered empathy – grasping circumstances intuitively and in a way an algorithm could not hope to achieve. It is a function of being made in the image of God and endowed with a conscience that informs richly. Computers remain a poor facsimile of humanity, good servants, but bad masters.
Mostly we have no idea where this is heading because it is an esoteric area of expertise belonging to a strange, unknown caste. Amid some tired political debates, we would do well to give attention to this one because, amid the undoubted opportunities, it presents a genuine risk to the cause of justice, equality and what it means to be human in a world where God has granted us great freedom.
Obama's Covert Wars
The use of drones is going to change warfare out of all recognition in the next decades.
Through A Glass Starkly
Images of traumatic incidents caught on mobile phone can be put to remarkable effect.
What Are British Values?
Is there a British identity and if so, what has shaped the values and institutions that form it?