Man vs Machine: 0–3

By Craig Dennis, taken from Pexels

AlphaGo took the third game in a series of five matches against Lee Sedol, beating the world’s top ranked player in the game of Go.

Watching the match unfold, even as I am not a player and only roughly understand the basic rules, was a little unsettling. The US commentator of the match noted that AlphaGo was very good at ‘capturing large areas’, a reference to the lower portion of the board where it overcame Lee Sedol’s best efforts to contest the area. He also noted at times that ‘the AI did not seem to be taking the easy way out’, a reference to two or three occasions in the game where it could have made moves that could have ended the game but did not do so. If the AI could have a personality you might even speculate that it was being ‘arrogant’, declaring that areas of the section were won even without completing the series of moves that would seal its victory.

As a casual observer I can’t help but have ambivalent feelings towards AI having progressed to such an advanced level. This only confirms my beliefs that I will witness in my lifetime what people call ‘The Third Revolution’, where disruptive changes in technology forever alter the face of the Earth as we know it. What is clear today is that AI has reached and surpassed human intellect; perhaps only in the game of Go, but who knows what else tomorrow? Go, with 10⁷⁶¹* possible game outcomes, is already one of the hardest ones to crack in a game where there is perfect information i.e. all the pieces are laid out in front for all to see. By comparison, I read elsewhere that chess only has about 10¹²⁰* possible moves.

If machines are coming for us, then I worry about my place in finance. A world where machines are doing the same jobs once reserved for humans is already here. The UK bank Lloyds has already announced its plans to retrench some of its bank tellers, to be replaced by a machine interface. I envision that very soon it will not be just the low-ranking tellers who will be out of jobs. Kensho, a startup founded by former Goldman Sachs employee Daniel Nadler, is threatening the jobs of thousands of researchers and traders on the buy and sell side. Touted as a “Warren and Siri for traders, analysts and investors”, Kensho is a revolutionary software that offers insights into the financial markets traditionally done by Wall Street’s analysts and traders. Using natural language inputs, big data and machine learning, it produces analytics at a speed and scale that would have previously required an army of analysts to accomplish.

Now that, to me, is absolutely exciting — and worrying at the same time. If computational prowess and development in AI continues at the current rate, how long will it be before AI starts challenging fund managers to beat them at their own game? How long until analysts become redundant, who will take weeks to plough through company financials that the AI can do in a few minutes? How long until traders become redundant, outmatched by price-searching algorithms trained to exceute on the best prices based on rules written by its programmers?

To the naysayers, that day will come later rather than sooner. But I would prefer to err on the side of caution and suggest that it could come sooner than we expected. Interestingly, it was just a little over 20 years ago when Gary Kasparov lost to Deep Blue, the supercomputer created by IBM. Going further back, the first modern computers appeared some time back in the 1930/1940s. What this means is that it only took 70–80 years for AI to progress to what we see today. The futurist part of me waits with bated breath to see how AI will evolve to potentially be used to solve more of mankind’s greatest problems, but the narrow, self-serving part of me wonders what lies ahead for the many workers who will emerge jobless in the foreseeable future.

*numbers taken and from Wikipedia, not known for absolute accuracy. The figures are debatable mainly because of the way people define it (“is it number of possible moves, or number of moves that are legal in the game?”)

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.