Data-driven policymaking: what future do we want to create?

Posted on: March 28, 2019 by Robin Zondag

In this blog I explore the future of artificial intelligence & analytics for policy makers. I state AI & analytics can make a difference in policymaking and execution as it is inductive for personalization of policy making. But, be cautious, the consequences of personalized policymaking contain new ethical considerations, with both pros and cons, for which we need public debate to determine the rules of this new game.

Data already impacting our lives

Personalized decision making is not new, already decisions based on Big Data are made in many fields. We see it working in commerce where customers receive tailored personalized offers and for personalized medicine, where treatments are customized for an individual patient, often resulting in better health outcomes. Nevertheless, today’s topic is: How should we handle the consequences of personalized policymaking?

Next step: personalized policymaking

In personalized policymaking, fine-grained policies apply to different kinds of people. Just like personalized medicine, effects are measured by applying the correct regime of data collection, decision models, testing and accounting for personal circumstances. The tricky part appears when we need to make a decision based on these personal circumstances and the resulting effects.

Right or wrong?

For sure personalized policymaking will allow society to achieve the targets of policymaking (e.g. less accidents or safer streets) more effectively, just like personalized medicine and personalized commercial offers achieve better results. However, the question is whether personalized policymaking is fair to individuals. Aren’t all men equal to the law?

This theoretic debate behind personalized policymaking can be summarized in two sides of the coin: ‘utilitary’ vs. ‘righteous’. The ‘utilitarian’ stream argues the overall gain (for example when balancing between two or twenty victims). But is it ‘righteous’ if one of the two victims is you? The two streams could not be more opposite.

Let’s take an example: ignoring a red stop sign. Is it wise to judge different people different on the same actions based on their characters and circumstances, like any teacher, judge or good manager would? With today’s technology it would be possible to personalize the sanctions on ignoring red stop signs. This will result in less people ignoring stop signs, and therefore less accidents and deaths caused by ignoring stop signs. But should we want this as a society?

Is it fair when the algorithm decides to send a mild fine to your home address, because you are a very cautious person (and so you must have missed this stop sign just once). Is it still fair when you receive the mild fine, while, your neighbor is sent to jail, as the algorithm defines he is never cautious when crossing the road and probably only learns this way.

By applying Big Data for policymaking and execution we, as a society, have the opportunity to apply a more tailormade approach improving the effectiveness of policies, to use data so that we can gain in welfare, even to save lives. For example, self-driving cars have the potential to save millions of lives by reducing the number of road accidents.

However, it is unlikely they will be faultless. Therefore, people will die based on decisions taken by these algorithms. What is preferred in this case: millions of people dying in traffic based on human misjudgments or hundreds of people due to algorithm mistakes? If you feel it is the latter, humans will need to start accepting decisions from algorithms. Even when they are occasionally wrong. What policy to apply?

Current experience is that humans raise the bar much higher for computers than for other humans. We might miss out on better living standards if we blindly continue this way. On the other side, human dignity needs to be protected too. A big brother taking decisions for citizens is a very unattractive vision. Finding the right balance is a task for us, people living in this day and age. It is upon our generation to determine how to use these new technologies for the good and define the right policies for it. A great challenge.

At least, I know one thing. It is not on me to judge what is fair and what not. What is needed is a public debate.

Special thanks to Carline Nauta for co-writing, Kimberly Fidom and Marcel van de Pol for reviewing.

Share this blog article

About Robin Zondag

Global head of AI labs
Robin Zondag is the Global Head of Atos AI Labs, responsible for leading the development of AI business as part of the Google Cloud & Atos strategic partnership. He works with customers in the AI labs of London, Paris and Dallas to design new AI use cases and define their AI journey. With Atos, Google and their ecosystem Robin creates next generation business AI services. His aim is to make AI real for all businesses. Previously Robin was a Consulting Partner leading Advanced Analytics globally and Digital Transformation in Benelux & the Nordics. Strongly believing in the disruptive power of digital Robin Zondag works with clients to change the way they interact with customers throughout the customer lifecycle, to adapt their business models and to transform the way they run their business operations. He executes his job by cooperating with clients, Atos colleagues, Atos associates, large and small partners, and universities.

Follow or contact Robin