NEW REPORT: Identifying Citizens’ Needs by Combining Artificial Intelligence (AI) and Collective Intelligence (CI)

Identifying Citizens’ Needs by Combining AI and CI

Around the world, public leaders are seeking new ways to better understand the needs of their citizens, and subsequently improve governance, and how we solve public problems. The approaches proposed toward changing public engagement tend to focus on leveraging two innovations. The first involves artificial intelligence (AI), which offers unprecedented abilities to quickly process vast quantities of data to deepen insights into public needs. The second is collective intelligence (CI), which provides means for tapping into the “wisdom of the crowd.” Both have strengths and weaknesses, but little is known on how the combination of both could address their weaknesses while radically transform how we meet public demands for more responsive governance.

Today, The GovLab is releasing a new report, Identifying Citizens’ Needs By Combining AI and CI, which seeks to identify and assess how institutions might responsibly experiment in how they engage with citizens by leveraging AI and CI together.

The report, authored by Stefaan G. Verhulst, Andrew J. Zahuranec, and Andrew Young, builds upon an initial examination of the intersection of AI and CI conducted in the context of the MacArthur Foundation Research Network on Opening Governance. 

The GovLab together with Nesta is also conducting a related research project, Designing for Crowd and Institution Collaboration, which will produce an additional set of case studies on government uses of CI from around the world. Those case studies and associated findings will be published in 2020. 

The report features five in-depth case studies and an overview of eight additional examples from around the world on how AI and CI together can help to: 

• Anticipate citizens’ needs and expectations through cognitive insights and process automation and pre-empt problems through improved forecasting and anticipation;
• Analyze large volumes of citizen data and feedback, such as identifying patterns in complaints;
• Allow public officials to create highly personalized campaigns and services; or
• Empower government service representatives to deliver relevant actions.

The case studies included in the study are: 

• The US Department of State’s Use of Insights.US – Using Artificial Intelligence and Collective Intelligence to Understand Citizen Needs and Expectations
• Camden’s Use of the Camden Resident Index – Using AI to Personalize and Scale Public Service Delivery
• POPVOX – Using AI for Constituent Engagement
• Métropole du Grand Paris’s Deployment of Assembl – Collecting Knowledge to Shape Policy Development
• Carrot Insights – Using AI and Gamification to Promote Public Wellness

The analysis yielded three key, cross-cutting takeaways that could inform future work in the space: 

• The Field Remains Nascent: Across all five case studies, there is significant ambition. While each project has a clear vision for how AI and CI can be used to improve governance, none has had a perfect implementation. Rather, difficulties in attracting an audience to the project and making the insights useful and understandable were common.
• Organizations Need to Put Greater Emphasis on Planning Activities before Launching an AI/CI initiative: Civic organizations often seem to struggle with launching AI and CI tools. Though it is important to acknowledge that there is no such thing as “perfect preparation” and that no civic tech project exists in a controlled environment, groups can at least minimize the potential for disruption by being systematic and methodological prior to launch. 
• Building Cross-Sector Partnerships Are Essential to Bridge Capacity and Resource Constraints; and be Successful: Finally, almost all the examples show the importance of cross-sector relationships. Often, the organizations hosting AI and CI tools do not have all the resources needed to be successful. Though forming relationships can be difficult, involving multiple actors helps fill gaps in resources and expertise, identify possible analytical blind spots, and better reach the intended audience.

Read the report here.