Robo-Advisory, Fintech and AI
Looking around the world, we can see a growing interest in the application of artificial intelligence, especially in financial services. Major fintech events like the Money20/20 US and the Hong Kong Fintech Week gave it the pole position, inviting machine learning experts and futurists to debate it before giant audiences. That said, the practical application of AI in fintech is not a dream of the future, but a present reality.
Robo-advisors, using artificial intelligence to analyze data and manage a user’s portfolio accordingly, have been around for a while in the financial industry. They promise easy accessibility and less need for human interaction, thus providing services at reduced costs. One step further and we enter the domain of trading bots. Such bots make actual business decisions based on previously accessed market data and gainings while circumventing human errors.
But how far have we come already? Let’s take a look.
The State of the A.I. Industry
As CB Insights stated, the volume of investments into AI-related companies will at least account for 1.8 billion US Dollars this year. That is more than ten times as much as in 2012. It should be noted that 40% of the sum goes to companies working on applications of artificial intelligence, which can prove relevant to the financial industry. The fields of fintech and insurance, business intelligence and analytics, commerce and sales CRM see special interes.
It is fair to say, that economic interest in A.I. has gone through the roof in recent years. Artificial intelligence has been a topic for scientific research since the 1950s. In the financial sector, robo-advisor software entered the market as early as 2008, in the aftermath of the financial crisis. Initially dominated by small startups, traditional financial institutions took over control of the development a few years later. They acquired the small companies, presenting the technology to a larger client base. Today, more than one-hundred companies provide robo-advisory services.
But before the term robo-advisor was even coined, financial advisors and stock market service providers have already worked digital. They used software to manage their customers’ portfolios and cope with the increasingly fast-paced capital market. Compared to how AI evolved after the turn of the century, these are just basic processes, as AI expert Juergen Schmidhuber has pointed out at the WIRED Money 2017convention. They were “Little tricks”, performed by systems running on pre-identified rules and simple machine learning mechanisms. The software takes over repetitive paperwork and record matching, making a stock advisor more productive.
A.I.-based programs created after the millennium follow a more sophisticated approach, now termed “deep learning”. State-of-the-art A.I. gathers information, recognizes the patterns within and thus improves on its analysis. They are, as Schmidhuber puts it, “prediction machines”. In the near future, machines will be able to make informed decisions based on substantiated conclusions. What’s more, artificial intelligence greatly outshines human capacities when it comes to data analytics. It sorts out information faster than the human mind could. Additionally, artificial intelligence is not inconsistent or biased in its analysis (at least if no-one tells it to be). Given the right data, it can get us straight to the decision, skipping the research.
What the Future Holds
A.I. evolution hasn’t stopped there, of course: It now goes straight to the result. Autonomous robo-traders already buy and sell stocks in line with the data they gathered earlier. As these systems are self-learning, they themselves as they go, rendering human input obsolete (or monkey input, that is). Google accomplished a recent breakthrough in deep learning via its self-learning DeepMind technology. Artificial intelligence could also help with compliance and security. AI could track the flow of money by evaluating huge datasets. Additionally, thanks to machine learning processes, it would become more accurate in detecting fraud with each operation.
In the future, human stock advisors might have to accept a reduction of their client base. Or at least, they might perform duties very different from those they perform now. Virtual assistants not unlike those we know from our mobile devices will take their place. Experts and observers of the financial industry see the benefits of such a development. Using robo-advisors will not only allow companies to cut costs, but also introduce users without experience to stock trading. This would make secure investments more common.
There is every indication, that this assumption will become reality, given that artificial intelligence is all around us already. We rely on the virtual assistants on our mobile devices to give advice on what soap to buy and what movie to watch. The same technology can be applied to financial services as well:
- Personal robo-traders would make investments for us. Doing so, they factor in business reports, comments by analysts and global movements in the markets. They weight them against our personal information, like pre-configured investment limits or risk tolerance.
- Intelligent Personal Financial Management tools would calculate the spending habits of its users. They would recommend saving strategies — or directly adjust saving contracts or investment portfolios.
- A visit to the bank would largely consist of speaking to a chat-bot with a human-like appearance. In the background a powerful AI software calculates all transactions of the bank branch and sends these info to the bot.
And this is only what conventional computer-powered AI will be able to accomplish. The next technological revolution in AI already looms on the horizon. Some experts like Jeff Hawkins of Numenta doubt the potential of deep learning. They state, that sheer computing power is insufficient to imitate human brain capacity. In consequence, they develop intelligent machines, inspired by human brain biology.
Robots and Risks
It is important to note that wide use of A.I.-powered fintech software doesn’t go without risks, however. Could the evolution of artificial intelligence it beyond the limits of human intelligence? Forward thinker Elon Musk sees this as one of the greatest threats for humanity. Facebook-founder Mark Zuckerberg couldn’t agree less.
Autonomous A.I. making decisions that their creators might not have intended is more than just a bug. And we have to see that bugs already can cause serious harm. The Knight Capital Group had to learn this the hard way, for example. An error caused them to lose millions of dollars. Security precautions and frequent testing, maintenance and iteration might minimize such problems.
However, it gets more complicated when it comes to the autonomous setup of self-learning artificial intelligence. So, any robo-advisor’s predictions will only be as accurate as the data provided by the client. Artificial intelligence is a toothless beast without sharp-toothed big data. It is perfectly possible that an AI jumps to the wrong conclusions if this data is somehow tampered with on purpose — a whole new dimension of fake news.
Another reason why many experts remain wary regarding the spreading of A.I. in the fintech sector points at the human side of the fence. Artificial intelligence is still a highly experimental field. It has to be seen, if the regulations that are in place will hold up when facing new technologies.
AI — The Rapid Disruptor
Certainly, it will all be a question of speed as well. Research like the one by Google mentioned above shows how quickly AI software can evolve by simply processing information. Finding regulatory responses is a slow and cumbersome task, easily outpaced by technological advancement.
Maybe your robo-advisor will someday make you richer than those snotty ATM machines. And it’s just one of the reasons to keep an eye on developments in A.I. research. As big data is one of the requirements for many software solutions based on artificial intelligence, innovative start-ups from this line of work have all the opportunities to engage in this interesting field.
And so have companies specialized in compliance or stock exchange. It’s up to us to make the best of AI.
Originally published at trimplement.com on December 2, 2017.