Written by Adam Jones, Guest Blogger on Thursday 18 February 2016
Last week, Google’s in house Artificial Intelligence department announced a breakthrough. Their software has learnt to play a game. Not just any game, but Go. It is thousands of years old, and considered by many to be the hardest board game in the world with more potential move combinations than you can shake a stick at and a requirement to see patterns and combinations and to play by ‘feel’.
The complexity of the game is one thing, but the real point here is that Google’s AI engine has learnt to play the game. The Google team hasn’t programmed in the ~2.082 × 10^170 (i.e., a 2 followed by 170 zeroes) playing combinations, or defined a series of boundaries within which the engine should operate. Instead, the system studied thousands of past games, played millions of games against itself and built its own long term plans and strategies.
So what has this got to do with financial services? Well the simple fact of the matter is that Google has already developed machine learning to such an extent that it can interpret the world around it and learn enough about the hardest game in the world to beat the European champion at it five times in a row. It won’t be long before they, and the many other firms developing the various strains of artificial intelligence (machine learning, natural language processing, visual perception etc.) begin to use their technology for substantial commercial operations.
We already see discrete applications of artificial intelligence in fund management. Investment firms have begun looking at how they can evolve their trading algorithms to take advantage of leading edge technology. Bridgewater, the world’s largest hedge fund manager, has set up its own artificial intelligence team to work on better ways of managing its $169 bn assets under management.
On a retail client level, it is also easy to imagine big improvements which could be made in a range of areas; the delivery of financial advice, investment selection, budget management and a raft of general insurance applications are just a few of the obvious examples. The challenge is to not allow our thinking to be constrained by existing processes and concepts. While financial services firms work on AI to improve their position in their relative markets, companies outside of financial services are using artificial intelligence to improve their customer propositions.
When viewed in this broader context, the work being done by Google, IBM, Microsoft and Facebook looks even more powerful. When you combine the capabilities being developed in artificial intelligence with the power of big data, even concepts so deep rooted as the divisions between financial services, healthcare, utility provision and retail will eventually be broken down, and the customer will be at the centre of it. I can’t help thinking that Google will end up accidently fixing a client’s financial services needs as a by-product of the pursuit of other goals.
Of course, we aren’t at that point yet. There are still a number of hurdles which need to be overcome before Google’s game playing AI engine is dominating the financial landscape, not least of which is the regulatory framework. If you thought it was a challenge trying to regulate bad behaviour out of a group of misguided bankers or IFA’s, try regulating something which is smarter than you and quicker than you. Day in day out, firms in the UK grapple with the various rulebooks and guidelines published by the regulator. We might soon witness artificial intelligence engines devouring these rulebooks and feeding back on the contradictions and inconsistencies within them. The regulator and the government need to be in a position to provide sensible and proportionate guidance to regulation for these firms, and to do so in a way which still manages to protect the UK retail investor and the UK economy.
We have talked before at Altus about the challenges of automating advice, and also the impact of the digital revolution on the workforce. While we have argued technological advancement isn’t that hard, perhaps Google have just shown us that it’s really child’s play.