Science and Policy Integration Bootcamp and the Action Learning Team on Artificial Intelligence

For a science-based department like NRCan, integrating science and policy is essential to better understand and address emerging and increasingly complex challenges. Over the past few years, we have seen innovative initiatives evolve to strengthen the integration of science and policy, as well as new approaches to mobilize the breadth of talent, skills, and perspectives across the science and policy disciplines of the Department.

NRCan’s annual Science and Policy Integration (SPI) Leadership Bootcamp is one example of this. Launched in 2016, this week-long intensive training encourages greater cross-disciplinary and cross-sectoral understanding between science and policy communities. It also provides participants with tools and networks to enable them to be leaders while contributing to the Department’s on-going efforts to strengthen evidence-informed decision-making.  Participants learn and see firsthand how science and policy integration is a “two-way street”: science informs policy while policy helps us determine our science needs.

image

Figure 1: 2017-18 Science and Policy Integration Bootcamp Participants

As part of the extended Bootcamp experience, participants from science and policy backgrounds form Action Learning Teams (ALTs) to advance a real NRCan priority over a six-month period, applying their newly developed skills and appreciation for science and policy integration. As the team conducts research and analysis, they are encouraged to consider a wide range of perspectives (e.g., the private, public sector, and beyond).  They are also encouraged to take risks, to be bold, agile and innovative. The ultimate outcome included presenting recommendations to senior management to address their science-policy challenge.

In 2017, the three themes included Indigenous Knowledge, Artificial Intelligence and Seafloor Mapping. This year, the Bootcamp is focussing on geospatial, resource extraction and advanced materials. As these diverse themes demonstrate, the Bootcamp provides the flexibility and agility in dealing with emerging issues and linking them to broader trends in science and policy.

Artificial Intelligence ALT as an exemplar of the Bootcamp model

For example, the “Artificial Intelligence (AI) ALT" was tasked to critically think about AI in a ‘foresight’ exercise, focusing on the long-term impacts of AI for the Department. However, midway through the process, the ALT’s mandate changed and the team needed to pivot to respond to the needs of the department. The renewed focus was to research the potential impacts of AI on the workforce of Canada’s natural resources sector and to identify ways for AI to help NRCan improve its services to Canadians.

The AI ALT quickly mobilized, drawing on their diverse skill sets and unique position as a science and policy integration team. The team ‘self-organized’ into two sub-groups to complete the tasks.  The AI ALT assigned an overarching project co-ordinator and two members volunteered to be co-leads for each of the two research tasks. Meeting chairs were rotated to increase inclusive leadership and regular teleconferences were held to ensure participation of regional office team members.

“I think this project really challenged me to get outside my comfort zone, for example presenting to senior management committees, and learning from people that I wouldn’t usually have the chance to work with.”
- Elizabeth Carmichael
Science and Technology Advisor

The fluidity of leadership, open communication and workload sharing helped to remove silos, and develop professional skills. For example, the most junior members were encouraged to present to senior management level committees to support their career development.

In the team’s efforts to seek new and innovative approaches, members sought input from several perspectives: NRCan sectors, other government departments, other levels of governments, research institutes, and the private sector. This extensive outreach enriched the group’s learning experience and collective knowledge related to AI, and certainly informed the path forward.

One research outcome included a small-scale pilot proposal to apply sentiment analysis to public engagement processes. This was a highly innovative approach compared to conventional applications previously used. Sentiment analysis could pave the way for future NRCan public engagement processes to become more agile, equipped and inclusive.  The efficiency of the AI tool to sort public comments in real time reduced analysts’ time devoted to repetitive tasks, improved engagement response rates, and opened up more time for critical thinking. To enhance inclusivity and the democratic process, the pilot project aimed to identify geographical areas with low-engagement rates - enabling the Department to ensure hard-to-reach Canadians would have equal opportunities to express their views on future natural resource development projects.

image

Figure 2: How can AI help NRCan better serve Canadians: Artificial Intelligence Action Learning Team findings

As part of our final recommendations for how NRCan can become AI ready, there were many links to Treasury Board’s work on AI and proposed principles in their ‘Responsible Artificial Intelligence in the Government of Canada’ white paper. Most notably, we focused on how AI systems can be deployed in a manner that minimizes negative impact to employees while improving service delivery. While AI can give NRCan new capabilities in our work and service delivery to Canadians, we need to remember that it is not a magic bullet solution and deploying AI in any given circumstance needs to be done with an ethical lens.

As with any project, there were some challenges along the way. This included connecting with team members from regional offices and labs. The team often joked about the irony that we often struggled to get video conferences to operate seamlessly even though we were dealing with a research topic as technologically advanced as Artificial Intelligence. These simple tasks had turned out to be a real challenge. Nevertheless, this was all part of a learning opportunity – one that paralleled the research – to make sure everyone was included, regardless of the distance.  Science and policy integration is undoubtedly a team effort – requiring continual enthusiasm, curiosity, collaboration, excellent communication and most importantly, a desire for change.