When computers and artificial intelligence were just manifesting in the world, Ray Dalio saw a future where these technologies could be harnessed to create better decision making processes. Dalio foresaw how automation and machine learning via algorithmic decision making could not only change the world of finance, but decisions made by everyone around the world.
What is algorithmic decision making?
For decades, financial institutions made decisions based on the intuition and knowledge of their traders and employees, who determined what financial decisions would serve the company best. Over the past decade this has slowly changed, and now, algorithms are driving financial decisions for some of the world’s biggest companies. Algorithmic decision making occurs when a computer (or network of computers) determines the most optimal decision based on previous decision data and a variety of inputs.
Ray Dalio explains how his firm has been working on algorithmic investing since the 1980s. In his model, human investors utilize automated systems which suggest investment decisions based on algorithms. At the end of the day, it is still the human’s decision on whether to make an investment.
The amazing thing about algorithms is that they can learn over time. As new information is uncovered an algorithm can be adjusted to reduce its flaws. In the world of investing, this helps to perfect an investment strategy over years or even decades. It is believed that because algorithms are unbiased, they are able to make decisions strictly based on data and facts, and without the emotional inputs that human beings utilize in decision making.
Ray Dalio sees an automated future
After a few years working in the finance industry, Ray Dalio started Bridgewater Associates in 1975. As an investment management firm, Bridgewater would go on to be the world’s largest hedge fund by 2005, managing $160 billion for its clients. It was during his years at Bridgewater where Dalio began to recognize how financial institutions could better make decisions with the help of artificial intelligence and algorithms.
Such algorithms would learn over time based on previous user input and previous outcomes to produce the most optimal decision making techniques. But Dalio didn’t think this concept could just be constrained to financial markets, instead, he recognized that algorithmic decision making has the potential to change the way we as human beings operate in the world.
As Dalio saw the power of algorithms in his business, he wondered whether the same principles of automation could be applied to other areas of his business. To test his theory, Bridgewater implemented a management coach system which takes inputs from employees to build an algorithmic model of how to manage the business. Employees rate each other’s performance, alert managers to employee dissatisfaction, and provide valuable data which informs the management algorithm as what staffing decisions to make. This model is often cited as one of the reasons why Bridgewater has been so successful in recent times.
However, management is just one areas in which algorithms are being used to inform decision making.
Algorithms are everywhere
Take safety as one example. Most people tend to believe that police should investigate potential crimes using not only the facts, but their human intuition as well. However, this idea may be based more on emotion and less on facts and evidence. Certain localities are now utilizing algorithms to determine the risk factor of children based on their home conditions, family life, and other inputs. In fact, early reports of predictive analytics used in Pittsburgh, PA show reason for optimism. Brett Drake, a professor in the Brown School of Social Work at Washington University in St. Louis professor Brett Drake, who studied these early tests is optimistic. “Given the early results from Pittsburgh, predictive analytics looks like one of the most exciting innovations in child protection in the last 20 years,” says Drake.
Meanwhile, in the mortgage lending industry, algorithms have been the norm for over a decade. Research from 2002 suggests, “[automated underwriting] systems more accurately predict default than manual underwriters do.” This is said to lead to increases in approval rates and accuracy for applicants who may have otherwise been missed.
In the legal world, it is estimated that algorithms would actually improve decision making in the court system. In New York City it was estimated that jailing could be reduced by 41.9% without any increase in crime by utilizing algorithms in bail decision making.
These are just a few examples of how algorithms are informing better decision making on a macro scale, creating better, more efficient institutions and business processes. Yet, the same principles that drive the aforementioned algorithms can also be a driver in the success of personal decision making as well.
Improving your everyday life
What Dalio has uncovered is a decision-making process that is not just applicable to financial decisions, but to most life decisions as well. This is because it is proven that as human being we are terrible at understanding ourselves and our decisions. We are likely to recall past results in a skewed manner and misremember historical events. This causes us to inform current decisions with false information.
These problems can all be solved by algorithms. As Dalio puts it, “ As these machines help us, they will learn about what we are like—what we value, what our strengths and weaknesses are—and they will be able to tailor the advice they give us by automatically seeking out the help of others who are strong where we are weak.”
His book, “Principles” describes how human beings are on the brink of utilizing algorithms for many of their daily decisions. Think of a scenario you have encountered many times before, whether it be at work or at home. Now imagine that instead of basing your decision on your emotions in the current moment or your skewed view of the past, you are able to access data of your previous decisions to inform your current situation. Additionally, your decision would be based on data of all people to ever make the same decision in the past, in effect crowdsourcing information to inform your current decision. This data would be completely unbiased and even improve in its predictive ability with each passing decision.
One of the easiest places to see algorithms at work is in digital content and shopping. It is estimated that 80% of decisions made on Netflix and about 33% of decisions made on Amazon are a direct result of algorithms which take a user’s viewing history and ratings across all users to determine what would be the best content for that particular viewer. These algorithms are constantly adjusting suggested content based on recent decisions by the user.
Such technology is now even predicting what you will write in your next email. Google has introduced Smart Compose, which predicts email word sequences in real-time. This makes it appear that Google knows what you are going to write before you even write it, and this is thanks to algorithms and machine learning which use your previous emails, as well as the emails of other Google users, to best determine what you will write in your current email.
What about biases?
If human beings have their own biases, wouldn’t these biases play out in the humans who design algorithms and just put human bias in algorithmic form? This is a question addressed in a paper by Songül Tolan, a researcher at the Joint Research Centre, European commission. Tolan notes that evidence suggests algorithms could even make things worse. “There is overwhelming evidence showing that algorithms can inherit or even perpetuate human biases in their decision making when they are based on data that contains biased human decisions,” Tolan says.
It is inevitable that at some point an algorithm will make a biased decision. When this happens many questions will follow. Who do we blame? Who is at fault? How is the decision rectified? What are the legal ramifications? Until there are standards on how to deal with biased algorithmic decision makers it will be difficult to implement these strategies throughout different areas of business and life.
This may be the reason why the public is still skeptical that algorithms will provide as much value in daily life that is currently being touted. According to a survey from the Pew Research Center, the majority of adults in the United States still believe that computer programs will always reflect the human beings that designed them. This is quite a skeptical stance on the automation and learning capabilities of decision making tools.
Algorithms are here to help
Decision making processes based on algorithms and machine learning still have a long way to go. Without eliminating biases in these technologies which are also inherent in human decision making, they won’t provide much more use than the emotional responses that currently drive human decision making.
However, it is likely that over time such biases are addressed and algorithmic decision making becomes more of the norm throughout the lives of all humans around the world. This isn’t something to be afraid of, instead, it should be celebrated. The work that Ray Dalio started decades ago is coming to make our lives more efficient and provide us with the tools to make the best decisions in our own lives.