Navigation path

Embedding digital science into future policy making

Connecting scientists with policy makers to build accurate digital models of the world and to perform detailed experiments and simulations of natural or social phenomena; foster global systems science approaches to effectively manage resources, including energy grids, transport systems, healthcare, water, food, etc; develop and implement narratives for change management.

 

Impact: 

This approach is potentially very impactful, as – perhaps for the first time in history – the growing demand for evidence-based policy making can be satisfied by means of abundant supply of intellectual labor and cheap computing. However, it is also risky: the flip side of cheap computing is that almost anything can be proved, by some definition of "proved". Interpreting results from advanced modeling requires a deeper-than-casual familiarity with modeling techniques, and some notions of probability theory and, increasingly, metastatistics. Since these are unrealistic to expect from policy makers, the risk is for policy makers to consult data analysts much in the same way that ancient rulers consulted augurs and wizards – and, consequently, to make their decision making completely unaccountable ("Apollo demands that we wage war on Thebes" or "models show that drone surveillance could reduce terrorist attacks by as much as 80%").

Does data science and the ethical and epistemological issues related to it bend democracy towards expert rule?

Plausibility: 

The plausibility of policy making integrating data science is high. That of doing it well (with the critical distance that comes from understanding the theory and having some ideas of the tools) is unfortunately low.

When big data are involved, interpretation of results can be really difficult. Through machine learning techniques, you can evolve classifying algorithms to make predictions; but that does not necessarily mean understanding why the algorithm predicts the way it does, even for the analyst. Machine learning is vulnerable to Black Swan events.

The cognitive load of fully taking on board advanced modeling as a tool to aid policy making can be high. This will not stop ICT vendors to push big data solutions, and policy makers to buy and use them. Unfortunately, not all uses will make sense.

Policy Type:

Impact: 

Plausibility: 

Support: 

Curators: 

Policy structure

Related events