We need to talk about Artificial Intelligence
By Georgeta Dragoiu and Tarig Hilal
Our 17th Century forbearers had little idea of the magnitude of what was coming when they embarked upon the industrial revolution. We have no such excuse.
Last week’s launch of the HBO series, Westworld, presented viewers with yet another dystopian vision of a future where Artificial Intelligence (AI) has gone awry. To coincide with the release, Sky Atlantic commissioned a survey of the British Public.
The survey found that 40% of those interviewed feared that robots could destroy humanity, prompting the Daily Express to run the headline “AI robots will take over the world and DESTROY human race”. Millions of Americans feel the same way.
Whilst sensationalist headlines may produce blockbusters and shift newspapers, they are distracting us from crucial conversations about the impact of AI in the present.
AI is a spectrum of technologies, imagined and real. On one extreme is the super computer, HAL from 2001 Space Odyssey, or Terminator. On the other are the task specific algorithms that are now part of our everyday lives; the Facebook newsfeed, Amazon’s “recommended for you” service, and now the technology behind autonomous vehicles.
These are already having a profound impact or will do so in the near term. Facebook’s newsfeed is one example of how narrow forms of AI are shaping society. Its algorithm’s promote topics posted by friends over publisher content creating echo chambers where views are affirmed more than challenged. Increasingly we are disconnected from those who hold different perspectives, an ironic outcome for a company whose stated purpose is connecting the world.
In other areas AI is raising concerns over its potential for discrimination. Recently, Propublica highlighted the dangers of using AI as a predictive tool for future criminality, suggesting that it may be resulting in a bias against black men. Amazonfresh recently came under fire for an algorithm that excluded poorer neighborhoods from its delivery service.
One of the revolutions that AI promises is in transportation. Uber has already launched its driverless car service in Pittsburgh, and driverless vehicles are poised to become mainstream within years. The benefits that these developments bring come with significant risks. The effects of automation are likely to be felt first in trucking, taxi and bus services, a move that could potentially put 2.6 million jobs at risk, a number equivalent to the job losses of 2008.
We used to talk about economic transformations in generational terms, now they are occurring within generations, in the future possibly multiple times. Just last year, Uber promised to create one million jobs worldwide; now the driverless car industry is setting itself up to take those away. Communities are already struggling to keep up with the rate of change, the future promises for that change to be faster still.
There is a view that all of this is inevitable, that people should simply adapt and if they don’t, then their suffering is a sad but necessary tribute to history’s long march. It is a view commonly held by people who do not expect to pay the price. But in the long run, the effects of AI will likely be felt by far more than Uber drivers. Already bots are writing newspaper articles; algorithms are encroaching on the work of bankers, lawyers, doctors, coders and maybe one-day politicians and industry leaders. Even if this possibility were not on the cards, it is incumbent upon us to assuage some of the more negative consequences of AI.
Once upon a time we might have cried ignorance. Our 17th century for forebears for example were for the most part unaware of the magnitude of the looming industrial revolution and anyway lacked the means to do anything about it. We no longer have this excuse.
This time though we know the revolution is coming and have the resources to consciously shape the future. We can draw upon the lessons of past social disruptions, government, experts, a robust democratic system and the vast resources of the companies that are driving this change.
It is critical however that every day citizens have their say. Public involvement is not just a question of principle, but pragmatism too, improving the quality of choices made and ensuring acceptance when they are. This has not happened yet.
The state of the political debate on both sides of the Atlantic is a case in point. In the US the leaked Trump tapes dominate the news cycle, and in the UK politicians still talk about the economy as if it were the 1980s, where the biggest question is whether or not rail should be nationally owned. All the while new technologies are transforming our societies.
Whilst it may not be Hollywood’s job to inform, the popular press surely has a duty to. Our politicians and thought leaders certainly do. It is time that the discussion around the implications of AI moves beyond popular entertainment and the policy fringe to the centre stage; from the film set to the town hall, from the corridors of power to the public square. We could start by making it part of the next Presidential debate, it would be a worthy respite from the recent ugliness. Afterwards we could all enjoy an episode of Westworld.
Tarig Hilal is an Associate at the Future Society at Harvard Kennedy School, an organisation committed to better understanding and discussing the profound consequences of the current technological explosion.