Many EU institutions are discovering the benefits of analytics to support policy decisions. But ultimately, no one has more potential to enrich output quality than policymakers themselves. Therefore, the entire organization must understand AI and analytics well.
We call this the democratization of AI, bringing together the ideas of data scientists and domain experts.
According to McKinsey, incorporating domain knowledge into analytics projects will lead to more informed results. If people understand AI and know how to ask the right questions, they are more likely to come up with innovative ideas that promote the development of new models. We see this in many industries and the same is true for policymaking. When domain experts rely on analytical knowledge, they can think differently and approach problems from different angles.
Better insights and feedback
How does policymaking benefit from analytics? Taking more information into consideration for decision-making leads to greater transparency, higher quality and a more robust feedback mechanism to test policy outcomes. Data and analytics can be key in defining a problem by modeling the context and simulating what might happen. Combining multiple data sources leads to better insights about individuals, entities and regions exposed to a particular risk. Early warnings will enable early intervention, risk mitigation and a more proactive approach.
Several EU stakeholders are working with SAS to take advantage of analytics in their policy decisions. To this end, they focus on how AI tools can leverage the functional expertise and domain knowledge widely available at the European Commission and Institutions.
Closing the gap
To pave the way for fairer and more effective policies, we must close the gap between data scientists and the consumers of the technology. Data and analytics will not automatically lead to better policy outcomes. If analytics models are fed with unrepresentative or manipulated data sets, the process can be undermined and ultimately have the opposite effect. We call this the “GIGO principle” – garbage in, garbage out. To avoid biased datasets, domain experts should be trained to support data scientists and validate their findings.
Yet there is another reason why we need to upskill domain experts and democratize AI. There is an alarming scarcity of data engineers and scientists in all industries. Especially in the public sector, attracting and keeping technical roles is becoming increasingly difficult. So, if you can upgrade the non-technical minds of domain experts and other people in the organization, you can accelerate the impact of analytics. Armed with powerful software, experts can perform detailed diagnostic analyses and work with data scientists to generate machine learning models that support new policies.
Platforms like SAS help demystify the process of analytical deployment, enabling domain experts to add data and produce insights without the direct involvement of a data scientist. The low code/no code approach and user-friendly interfaces allow more people without a technical background to use analytics. Especially for data preparation, data understanding, modeling and evaluation of models, domain experts have a greater role to play. This will ultimately result in a more hybrid approach to complex cases and help alleviate the current tightness in the labor market.
