It seems a very current obsession, but the truth is that artificial intelligence both as a term and an area of research has been around since the 1950s. In the past six decades, we were promised the impending arrival of robots that would fold our laundry perfectly, machines that could automatically translate languages with great aplomb and luxurious flying cars that would jet us about on their own, as in the famous American cartoon show, The Jetsons. But the task of building synthetic cognitive capacity was not a trivial one. It has taken decades for AI to mature, and these past visions of a possible future came to be seen as mere science-fiction scripts written by optimistic boffins who were, perhaps, promising a bit too much.
But now, things are changing.
In its third coming, artificial intelligence has finally emerged from the long AI winters of the past when the technology became anathema for investors and business decision makers. Through the successes of Waymo’s self-driving cars, Amazon’s Alexa personal assistant, SparkCognition’s predictive maintenance AI and SenseTime’s somewhat frightening facial recognition technology, we now know that AI is no longer an impossible aspiration. To quote President Putin, it had been transformed into an instrument through which to “rule the world”.
What a reversal of fortunes this has proven to be for AI researchers! One only needs to peruse a current newspaper to be entirely inundated with stories of the latest AI developments. China and France, are launching national AI policies with a view to gaining the upper hand in developing the technology. The United Arab Emirates just announced the appointment of the world’s first AI Minister, Sheikh Omar bin Sultan Al-Olama. Even in the United States, the birthplace of artificial intelligence and the world’s leading AI power, myriad efforts are underway to determine how best to win what has been widely characterized as “the AI race”.
My own work with the US military, think-tanks such as DC-based Center for a New American Security (CNAS) and in founding and running one of the fastest growing American AI companies, SparkCognition, has made it clear that there is much we must prepare our societies, militaries and indeed, ourselves, for.
But above all, there is the economy.
The confluence of artificial intelligence and robotics will most certainly have an indelible impact on the jobs of today. We mechanised human muscle centuries ago, and when we pair it with a synthetic mind, nearly every form of economic labour can be automated. While AI’s near term impact on jobs is generally forecast to be modest, it is the medium term we must brace for.
PriceWaterhouse estimate that two per cent of UK jobs will be impacted by the early 2020s. But this effect will gather steam very quickly. By the mid-2030s, the same report estimates up to 30 per cent of jobs in the UK will be automated.
And just what kinds of jobs will go the way of the machines? White or blue collar? Low-wage or high-wage? The reality is that the impact of AI will be felt across the board. In a study of average compensation across different British jobs, the Independent ranked a variety of careers by their average weekly earnings. Pilots and aircraft engineers, earning an average of £1,800 were at the top of the list, with CEOs a close second (£1,580.70). Meanwhile call center workers (£326), valets (£309), waiters (£266) and nursery nurses (£295) were among the least paid. If we analyze current AI and robotics capabilities and extrapolate future developments even in conservative ways, there is little doubt that much of what an aircraft pilot does can be automated. This may not mean that we will have pilot-less civilian aircraft in a few years, but it will imply a need for fewer pilots for a larger number of flights. With rapid advances in text-to-speech applications, natural language processing, robotic process automation (RPA) and voice recognition, call center workers do indeed have a lot to be concerned about. And perhaps with autonomous cars, valets will see their numbers dwindle too.
Jobs with two characteristics do stand out, however, as resilient even in the face of automation.
First, jobs that require the “human touch”. These are caregiving jobs such as the nursery nurse cited above. At the end of the day, we humans perceive the world via a series of subjective experiences. And in certain contexts, we want to feel the warmth of a human connection. It makes us feel “better”. Jobs that primarily deliver this sort of value are beyond the capabilities of machines far into the foreseeable future.
Second, jobs that we as a society agree have value simply because they were done by humans. For example, a robot yielding a fine brush may produce an oil painting of a scene in a nearly photorealistic way. One could argue that no human would be able to capture it with such precision or detail. And indeed, while we marvel at the beauty with which Rembrandt captured the light and shadow in his “Storm on the Sea of Galilee”, ultimately it is not the brush work, but the idea expressed by a human mind manifest on canvas, that we find valuable. It is not about the number of pixels, or the precision. It is about a confirmation that we are capable of this… this painting, sonnet, sprint or athletic manoeuver. We will always find this confirmation, provided by artists, sportspeople, singers, sculptors and musicians to be valuable.
But is this enough? If 30 per cent of the country’s jobs will be automated by the mid 2030s, is there any solace to find in the fact that a few of us will still be paid to care for children or paint a masterwork, if we are blessed with the ability?
Clearly, not. As I discovered in writing, The Sentient Machine: the coming age of artificial intelligence, the dawn of widespread AI is not just about technological innovation and scientific progress. It is also an opportunity to understand ourselves better. The advent of AI can be a mirror unto humanity; why are we valuable when machines can outperform us?
I feel we, as sentient beings able to perceive both abstract ideas and tangible reality, possess intrinsic value. And no, this value does not accrue from our ability to perform labour, but is due to the unique perspective we bring forth. Machines can be monumentally faster than us, but pitted against the infinite space of ideas, they can uncover just about as much as we can… an infinitesimally small piece. No speed can conquer infinity. Thus, it is not speed, but perspective that truly matters.
Consider that economic value – beyond the obvious needs of the flesh – is a consequence of human agreement. We agree that famous musicians and cricketers should make a lot of money. There is no fundamental law of physics that says this should be so. Seen this way, value seems arbitrary, but it is precisely this arbitrariness that frees us to adapt our economy to the coming age of artificial intelligence. Whether through universal basic income, a robot tax, or by paying every citizen a “rent” on the use of their portion of Earth, we must find ways to move beyond today’s definitions of jobs and value.
This is not just technological innovation, but policy and political innovation. And to realise it we must engage our governments and policy makers, who, for the most part have been left behind by the march of technological progress. We must push them to develop the legislative frameworks through which we can prepare our societies and economies for a future in which citizens are able to reap the rewards of AI, and not merely be transformed into its casualties.