Soon, banks won’t sell to consumers. They’ll sell to bots, using bots, overseen by bots. Jersey is well placed to influence the law-making that will enable this.
AI will change finance. It will change the way that companies work – but far more importantly, it will change the way that the financial services industry and its markets work.
And it is not only technologists who envisage a future that combines fintech and regtech to create a better financial sector. The Bank of England’s Working Paper No, 674 on machine learning at central banks, published in September 2017, explored banking supervision and came to similar conclusions. AI is not only a fintech that can help individual organisations shift their cost-benefit ratios around products and services, but also a regtech that can help jurisdictions create better financial services sectors.
Digital Jersey are have already started to prepare the ground. The AI retreat it helped to host last September was a first step in helping businesses, regulators and the Jersey government to make sure the island can reap the benefits of an AI revolution that is, frankly, inevitable and of such a magnitude that we cannot ignore it a moment longer. John Cryan, the Deutsche Bank CEO, was famously quoted in the Financial Times (FT) last year saying that his bank is going to shift from employing people to act like robots to employing robots to act like people. Jersey cannot hold back this tide, so it needs to start planning to take advantage of the new opportunities that will come with the new technology.
What opportunities will these be? I think it’s time for creative thinking. When bankers talk about AI, they tend to think in terms of robo-advisers and chatbots, focusing on the use of AI by their institutions to either cut costs or deliver new services. Most banking investments in AI are currently investments in machine learning, by and large targeting fraud. But there are other more radical ways to use AI. As I asked at the Digital Jersey annual review in an echo of Fred Schwed’s 1940s financial services classic… where are the customers’ bots?
Customers will have access to AI just as powerful as the banks’. Retail customers’ smartphones and fund managers’ tablets will connect them, permanently, to an intelligence far greater than their own. This cannot happen soon enough for me. I’m not smart enough to choose the right credit card, pension or car loan, so clearly, I’m going to want my bot to take care of things for me. But which one? Should I choose the Virgin Money bot, the best-performing bot over the past 12 months or the Google self-taught super intelligent bot that is also the world Go champion? (The new Google Go bot that taught itself to play has already beaten the old Google Go bot that was taught to play by humans — by a hundred games to zero.)
I’m not sure I really want to be in the loop for a discussion about pension plans or insurance, but I do want some sort of confidence that there’s a regulator in the loop, and that, if push comes to shove, my bot will be held to account to explain why it made the decisions it dis. Regulated bots are the future.
You can see why this is inevitable. My regulated bit is going to negotiate with the bots of the regulated financial institutions in order to obtain the best product for me. It’s not going to be swayed by the logo or the branch design or a double-page spread in a newspaper. It’s going to be swayed by price and performance. So if a bank is trying to sell me a mortgage or a credit card, it is wasting its time showing me gibberish advertisements involving astronauts riding horses, or whatever the last one I saw was. Brand. as the industrial age’s mass-market substitute for incomplete information, becomes meaningless.
Perhaps the regulators will have a list of ‘authorised’ bots, much as they have a list of authorised financial advisors now. But what is the super-intelligent learning bot outflanks the regulator’s super-intelligent learning bot? What if the bot makes the regulator think that it is acting in my best interests, while it’s actually acting in the best interests of the ruling political party or Mark Zuckerberg? As economist Diane Coyle pointed out in FT last year, it may be that transparency is the key to making this work, which flags up at least one area where the technology of shared ledgers and machine learning, blockchains and bots, may come together.
My head hurts. How the banks’ bots will interact with the customers’ authorised robo-representatives is at the same time fascinating and frightening – not for customers, but for the banks unable to deliver the cost-effective products and services that those representatives will demand. This will be a world where the bank’s AI sells to my AI under the watchful eye of the regulator’s AI.
If I am right and AI is an event horizon for the financial services industry, then even though we cannot see (or even imagine) the other side of the introduction of true AI into our businesses, we can see that our traditional ‘laws’ of cost-benefit analysis, compliance and competition will not hold. So it is important to start thinking about what the new ‘laws’ might be and how our financial services can help formulae them, thereby stepping up to the plate for the jurisdictional competition that is certain to emerge.
It’s no wonder Jersey Finance chose to focus on AI at its annual private wealth conference in London this year. If private wealth practitioners are not already talking about the impact of AI on wealth management solutions, they soon will be.
My view, and the view of Digital Jersey, os that a financial services sector that is more efficient (cheaper to operated), more productive (more desirable for businesses) and more effective (a beacon to policymakers and regulators) is an entirely reasonable vision for Jersey. And it is an attainable goal if we start working towards it here.