Will computers beat us at our own game?
What if you were the European champion in the world’s most complex board game, and were beaten out by a computer? This is exactly what Fan Hui, European champion in the game Go is experiencing right now.
Announced last week, a software program called AlphaGo designed by Google’s artificial intelligence team has become proficient enough to outsmart the European champion in Go not once, but five times in a row.
This breakthrough is a potential game changer for the artificial intelligence (AI) industry, as well as the economy overall.
Why beating Go is a milestone
At first glance, Go can look similar to chess, but is vastly different. It is played on a 19x19 grid with the simple objective of capturing more territory than your opponent by placing coloured stones onto the board. On any given turn, there may be more than 250 possible moves. Overall, there are more board layouts in Go than atoms in the universe! It has long been thought that no computer program could replace a human brain in solving this complex game. Not true anymore.
In order to beat the European champion, AlphaGo had to learn how to act intuitively and think like a human. To master this ability, the DeepMind team fed millions of human moves into the computer and allowed it to identify patterns in the play that led to positive and negative outcomes. This process was followed by the computer playing millions of games against itself and constantly improving its own decision making.
By combining these learning techniques with computational power, AlphaGo appears to have exceeded the skill level of the best human players.
This advancement is significant – it indicates that computers are no longer just for crunching numbers, but shows their ability to think creatively and develop solutions.
What’s more, technological advancements of this kind could very quickly shift from games to office tasks. Many tasks conducted by knowledge workers follow similar patterns on a day-to-day basis, with only simple details changing each time.
Many fields will be advanced by AI developments:
• Doctors aided by sophisticated computer software could more accurately analyze test results or MRI scans.
• AI could also bring personal assistants into the workplace, accompanying employees throughout their careers.
• An AI personal assistant might have the ability to learn and adapt alongside their human counterpart, significantly benefiting their life by amplifying their work and knowledge. We already see semblances of this through the use of intelligent personal assistant applications like Siri and Google Now.
Unfortunately, this also means the breadth of occupations up for replacement just grew. Computers already conduct the majority of trades in financial markets and can even write articles about sports and the weather, without the assistance of a human.
Oxford University has suggested that 47 percent of occupations in the US are at high risk of automation. Jobs in office administration, sales, services, manufacturing, and transportation face the greatest immediate risk.
On the one hand, with every major technological breakthrough (i.e., the creation of the steam engine, the printing press, and the motor vehicle) there has been both worker displacement and the creation of new jobs. Can we hope that with these recent advancements, there will be new jobs to meet demand?
We may not be so lucky. Workers may be permanently replaced by software programs that can run on any desktop computer. It may be that only the highly educated will be able to find work in the future; all other jobs will be automated at a lower cost than employing a human being.
At the University of Toronto’s recent conference discussing AI, Stanford’s Jerry Kaplan remarked, “the real problem is that the wealthy will need few, if any, people to work for them at all.”
Despite the fear of replacement, there is room for optimism. Even with sweeping changes in technology over the past few decades, the structure of employment has remained steady. Economies have displayed a dogged ability to bounce back from change, each time with unemployment rates returning to normal.
We face both a bleak and optimistic future with AI developments. Yet what the future holds for work and workers is anybody’s game.