Artificial intelligence software at the end of the rainbow

The incoming artificial intelligence software revolution is affecting the finance sector. It's not easy to draw the boundaries of such a broad concept as the financial sector, but evolving customer expectations, crypto-currencies, peer-to-peer landing and changes in the regulatory landscape are all affecting the way finance is executed today. And those changes are rapid. Enjoy this introductory video provided by the Financial Times. It suggests a few questions, we will take care of the answers in this post.

Artificial intelligence software and personal finance

Artificial intelligence

As I work in finance, people always ask me if we will benefit or be damaged by this kind of financial innovation. To answer the question, we have to think what finance really means. To me finance is about managing and transferring risk. When I buy insurance, I transfer some risk to the insurer. When I buy a stock, I accept the risk of default of the issuer and hope for dividends and increase in value of my title. When I buy a derivative I want to cover the risk of adverse variances of interests, exchange rates etc... So if artificial intelligence will help me to manage properly my level of risk, I benefit of it. The idea of companies using artificial intelligence software to make our life easier in the way we manage our money is exciting. There are already startups which build machines to help consumers make smarter decisions about their money. The market today already offers a set of service to manage personal finances, like Mint, Quick Book or Turbo Tax, but artificial intelligence is something different. It's about providing personalized suggestions based on big data collected from several sources. From your family status to the tax declaration through your shopping habits or any action on social networks, AI should be able to figure out your risk (adversity) profile and suggest a proper financial management behavior. This behavior is a lot more than taxes. You pay taxes once a year, you take many decisions every day which impact your financial status. For example when you decide to go out for dinner (and you probably state in on Facebook), you then refuel your car to go there (and you pay with your credit card), invite your girlfriend (and pay for two) and actually show up at that fancy restaurant (where you obviously check in with Foursquare) which you finally rate the day after on TripAdvisor.

It's not only about how many savings you're doing or if you're overspending; an Excel spreadsheet would be enough. An artificial intelligence software potentially can analyze a lot of behavioral information (where you live, what you eat, which sport you practice, how many cigarettes you smoke etc...) and for example predict your propensity to certain diseases and then suggest the right insurance for you. It will suggest investments. It will scan and understand all the financial products on the market and will suggest where to open a mortgage or simply the best bank account or utility provider offer in circulation. The more variables an artificial intelligence can manage, the better the way it can suggest a proper financial behavior. Would it be alarming or invasive? I don't think so as long as you can disregard its suggestions. When I wrote about artificial intelligence avatars living in our smartphone or in giant screens attached to the wall, I was thinking about this kind of services.

Artificial intelligence software and investor finance

But finance is not made only by individuals and private companies, it's made also by private and public institutions pursuing different scopes. Pension funds, banks, insurances, governments, hedge funds, investment funds and many others. Their interests range from speculation to maximization of benefits for their investors. And here is the bulk of the market. That's why there is a lot of buzz-word about computer assisted trading or even artificial intelligence algorithms replacing completely human traders. If you think it's all about speculation, you're wrong. Financial markets exist to allow the demand to meet the offer, they link people who have surplus money with groups of people who need it, help companies to raise funds, allow stakeholders to monitor trends, have a set of rules which protect the parties etc.

Today a growing portion of trading is managed by machines. Algorithmic trading is designed to react extremely quickly to market changes. The algorithms seek out and exploit small windows of trading opportunity, often measured in minute fractions of a second. Humans develop a mathematical model to identify trading opportunities, some companies have given traders the ability to code their own algorithms and to change them when market conditions change. But the game changer is artificial intelligence, software that can teach itself to adapt to changing market conditions, without any guidance or instruction from humans. Once launched is not updated by hand anymore and it recognizes patterns in the data crunched to make predictions about the market, which it translates into buy and sell order, all without any direct human involvement. Artificial intelligence software can analyze tens of thousands of different predictive patterns, something that no human can do and change (in seconds) the algorithms based on the market conditions.

So where are the benefits? Lower costs for the investors. Managing a portfolio has a cost, if it is managed by a self-learning machine, the cost will dramatically drop down. Supporters say computers boost liquidity, helping would-be buyers and sellers find one another without middlemen. So it's about opportunities and cost again (an intermediary has a cost by definition). I think that the artificial intelligence algorithms will go mass market at some point: I cannot afford hiring a personal trader to manage my savings, but I can probably pay a software if the cost is acceptable.

And what are the risks? As intelligent machines will replace largely human activities, human traders will be the first casualties of artificial intelligence in finance. But the biggest risk is about errors and losing control. In 2012 an important US market maker lost over $400m in 30 minutes because of a computer failure. The New York Stock Exchange had been already halted following a software problem. In fact, Securities and Exchange Commission is looking for ways to regulate algorithmic trading and has already introduced a set of rules to mitigate unwanted volatility.

But the worst risk is that the increasing automation of trading will marginalize returns, as strategies become more based on big data and computer technology. Now a machine can beat a human and react much faster, but when the battle will be all among different software the differences will be much smaller. When all the software will access the same data, what will make the difference again? I bet it has two arms and two legs...

A quick note on artificial intelligence and banks

Banks are not going to disappear. I agree many of us might feel some resentment towards the banking system, but its role is still important. This is not the place for a finance lesson, but the role of a bank is to collect cash from those having a surplus and lend money to those in need. Unfortunately banks today do not have a clue on how to lend money efficiently and protect their investment. If you think I've been too harsh in this statement I repeat it, they are almost unable to understand if the applicant deserves a loan or not. They spend a fortune in marketing to attract new clients, which they hardly know and as they are not able to lend money efficiently they push speculative investments to recover profitability. Initiatives like crypto currencies, peer-to-peer lending, decreasing rates and new regulations are undermining the banks profits. Their safety net is about data and innovation. The same artificial intelligence software which will marginalize returns and cancel another relevant portion of bank's turnover, will help them to identify where to allocate money. If it works properly we can expect the complete inversion of the paradigm. It won't be the entrepreneur to show up and ask for money, it will be the bank to chase the best mind with the best project and offer a loan.

Previous     Technology     Next