Mary Shelley’s Frankenstein , published in 1818, is considered the first work of science fiction. The protagonist, Victor, might be known today as the first Research Singularity Scientist. Halfway through the book, the monster shouts to Victor, ‘You are my creator, but I am your master; Obey!’. The million-dollar question is whether this situation can become reality 200 years later. Not with a monster, sewn together from the body parts of dead people, but with slick and shiny robots that have better brains than we have.
The development of AI is one of the most exciting trends around. Frankenstein is not science fiction anymore.
Arjan van Os Head of Innovation Centre
Artificial Intelligence (AI) is one of the most controversial topics in the field of robotics. Ray Kurzweil believes that, by 2029, computers will be able to do all the things that humans do. Only better. Critics disagree, but the possibility continues to get closer.
No one can deny that the groundwork is there already. The amount of available data has increased to zettabytes. More and more, this ocean of data is well structured and available to us through a growing multitude of APIs. And, of course, there is the ever-increasing power of applied computing. Mix these three together in smart algorithms, and they produce models and patterns in real time that humans cannot create or comprehend. It’s easy to see how close we are to creating algorithms that surpass human capability.
Introducing the monster
The development of AI is one of the most exciting trends around. Frankenstein is not science fiction anymore. The Big Techs invest billions in AI Systems, e.g. IBM Watson, which understands the nuances of human language and the context of a question. Consider also Apple’s ubiquitous knowledge navigator Siri or smart region-specific Cortana from Microsoft. They advanced from descriptive systems that produce information, into predictive systems producing insights and enabling better decision-making, to pre-emptive systems that are intelligent and learn on their own.
Embracing the change
In FinTech, the number of start-ups focusing on AI is rapidly increasing. Engineers and big-data scientists create new companies to offer advanced machine learning solutions for investment banking and cybersecurity solutions. But also in the field of Risk Management, the core function of every bank, predictive and prescriptive systems are introduced to produce more efficient risk models than the current more descriptive models.
In the next few years, we’ll probably see AI expand in Finance from back-end functions like overseeing trading decisions, detecting fraud, or risk analyses to the front-end functions like personal advice and much smarter ways to interact with a bank. For example, it could replace cumbersome questionnaires with natural dialogue. The ‘intelligent agent’ will offer better and faster personal service, 24 hours a day. It will help you with your insurance or investments, it will think along with you to decide which mortgage best suits your financial future, and it will alert you when something out of the ordinary occurs or may occur.
Preparing for the future
In the advanced version of this idea, clients – with the help of their personal banker avatar – will determine how their finances should be arranged, how much must be saved for the summer holiday this year, what the minimal balance on the household account should be, etc. The system will manage accounts and alert clients only when needed. Of course, it will also categorise transactions intelligently, provide feedback about financial health, and report on whether financial objectives are on track. These are all existing Personal Finance Management functions. However, in an AI reality, they will be more autonomous, user- and context specific, and will trigger clients with more intelligent insights and advice.
Providing the transparency
It will be interesting to see the acceleration of AI functionality in a context in which the new EU data protection regulation will be implemented. It is very likely that, by law, somehow the logic behind decisions and advice will have to be explained to the user. The robot advisor will have to be able to make the rationale behind his reasoning transparent. So apart from the ’how’ or ‘what’ it has to be able to answer, we can also ask a lot of ‘why’ questions to figure out the principles, hypotheses, rationale, logic, or thinking our robot advisor employed.
Of course, data protection and client security are also top priorities. We must explore all the ways to keep client data secure, while also making the most of the robots’ extreme processing power.
Remembering the client
ABN AMRO won’t be replacing human financial advisors with robotic ones tomorrow, but the concept is not as far-sighted as we might believe. And just how far AI will be adopted in the financial sector is as yet unknown. Will it remain a back-end tool that assists human advisors with providing ever-more accurate advice and projections? Or will clients embrace their financial avatars for their accuracy and availability? In the end, as always, those answers will depend on the client, and his or her willingness to adapt to the change.
But, by exploring and experimenting with this technology now, and preparing for the changes to come, ABN AMRO is doing the best thing we can do for our clients: learning everything we need to know to make the most informed choices when the time comes. As the creators of these sleek and shiny monsters, we have the best chance of preventing them from becoming our masters.