Seven easy steps to integrate AI in Collaborative Social Learning for “solving” Wicked Problems: A reflection.
If you are an educator or a learning designer in tertiary or university settings, you know there is something wicked about how students learn with group work. Many educators rely on group work to nudge students to learn not just about their subject domain but also about life-long competencies that can be used throughout their professional lives. However, group work can be a waste of effort if students treat it as an individual task that someone amalgamates into a barely coherent outcome the night before the presentation. Whereas some blame students, there is a deeper issue about the lack of effort in designing and reinventing meaningful student group work assignments.
I am currently participating in the “Open Networked Learning” course (ONL241), an open online course that explores the increasing use of digital tools and online collaboration for teaching and learning. Over the past weeks, my group and I have been collaborating to develop a course that leverages AI as a catalyst for social learning, with the aim of invigorating group dynamics. This reflection underscores the potential of integrating AI as a tool that meaningfully fosters social learning and collaboration in university-level group work.
We recognized that group projects are often underutilized and poorly structured, resulting in a scenario where the group work is divided among individuals who work in isolation and then barely put together an outcome. This individualistic approach typically hampers social learning because the task does not require meaningful and sustained student collaboration. A significant issue is the potential for freeloaders to undermine the course, disrupting the learning experience for the entire group. This is where integrating inquiry-based projects built upon an open, collaborative platform can change group dynamics.
How to make groups more meaningful?
Use wicked problems and AI to foster Collaborative Learning
We designed a group project based on the following three concepts: (1) Collaborative and social learning, (2) Learning by inquiry of “wicked problems,” and (3) Using AI to turbocharge collaboration by adopting multiple roles through the collaborative inquiring process.
Wicked Problems are good settings for inquiry.
We designed the course around “wicked problems,” which are complex problems that are often intractable and unsolvable. These problems are difficult to define, to begin with, because they contain uncertainty, ambiguity, and conflicting values and interests. Some examples of wicked problems include climate change, poverty, and crisis management. These problems cannot be solved by one person or organization alone. Wicked problems are often at the core of design thinking.
We chose to base our course on a specific wicked problem that aligns with the unique capabilities and competencies of our diverse group members: a microbiologist, a climate scientist, a business researcher, a fashion designer, and a social work lecturer. This decision underscores the crucial role of a multidisciplinary approach in tackling complex issues. The scenario we chose is the following:
“Although the industry has several initiatives to reduce its footprint and it has advertising campaigns touting recyclable garments, ‘Fast Fashion’ remains one of the worst polluters. Apply a multidisciplinary approach to create a workable solution.”
Then, we worked together to use AI throughout the inquiry process to make the collaboration meaningful. We organized the activities as airport books do, in the “seven easy steps” that simultaneously overpromise and underserve results. In our case, the seven easy steps direct attention to the paradox of the collaborative learning process as the steps are rarely seven and never easy. The paradox is even more acute because seven easy steps can never solve a wicked problem. However, the seven-easy-steps framework is useful for introducing a helicopter view of meaningfully including AI in class because each step is ontologically distinct. Here are the seven easy steps to include AI in inquiry-based social learning.
1. Set up a Wicked Problem (Meaningful, intractable) for students.
The teacher selects a meaningful and challenging wicked problem that is suitable for the audience. Then set up the hook by making the problem engaging for the students. The hook needs to be aligned to the topic of the class, i.e. engineers need to be hooked in by some engineering aspect, whereas law students need a hook that relates to law. Encourage students to identify personal connections to the problem, fostering empathy and motivation. Show why this problem is important for the world and very difficult to solve. After all, the wicked problem is (currently) intractable and utterly complex. Therefore, get students to define what aspect of the wicked problem they will address.
2. Team building and developing social learning skills.
The second step aims to build scaffolding for complex tasks. Building upon Gilly Salmon’s scaffolding approach to build the baseline for the team to work, students learn how to build connections between people with different competencies and backgrounds.
In practice, students talk about the wicked problem as if they were speed-dating. This practice allows students to address the problem in a less structured way for more open-minded ideas. The emphasis is on brainstorming and cooperation, not competition. At this point, the emphasis of the exercise is at the level of personal opinions and anecdotes. Therefore, students should frame the wicked problem through their professional expertise and discuss it with the other course members with different backgrounds.
AI comes into play by letting students play with AI chatbots to explore preliminary answers and frame the problem. However, it is important to create a paper trail of their discussion and interactions with AI, including which prompts they used and how precisely they used AI tools.
3. Use AI tools to problematize what aspect of the wicked issue students will address
Encourage the groups to interact with AI to find multiple angles and propose solutions to the Wicked problem. Using AI’s chatbot functionality can help students learn about the problems associated with the wicked problem and discuss them with other course members. If students do not have previous expertise in the wicked problem, AI can help them find a way to get into the problem and find connections with their expertise. They could also use AI to select/find experts to define different sides of the wicked problem.
AI is a tool, not an answer. Focus on the Chatbot functionality to use AI to ask meaningful questions and prompts about the nature of the problem. Use AI tools to analyze vast amounts of data related to the wicked problem in order to identify specific focus areas.
Students will use AI-generated answers as the baseline and then fact-check them with secondary data. The point of this step is to stop using AI as a knowledge machine and learn to identify its limitations and failures.
4. Introduce AI-enhanced Inquiry methods
The purpose of this step is to learn how to conduct a formal inquiry process using AI tools. For example, use AI-assisted research tools to analyze academic literature, industry reports, and documentaries focusing on specific aspects of the problem using Ai, but approaching the outputs with critical thinking and personal opinions. AI can provide help with data analysis depending on the wicked problem (e.g. large data sets). The idea is to learn to use AI in ways that are more complex than the chatbot. For example prompt engineering - learn how to write prompts for the chatbot or analysis of social media data using LLMs.
5. Integrate social learning and reflection into the inquiry
At this step, students should be able to establish their position addressing the wicked problem. Can students find a workable solution to the wicked problem? if so, can they defend it? Focus on the result of the inquiry problem in which students co-create new knowledge as a group. Students should reflect upon the wicked problem and what solutions they think are viable so that they can take a position on the issue, namely by proposing or endorsing a workable solution.
Students should reflect regularly on the ways AI supports and influences their thinking, discussing the trade-offs of relying too much on AI tools. One essential part of the process is creating a document trail for how AI has been used.
6. Integrate social learning and reflection into the inquiry
Even if there is no solution to the problem, students can still take a position and defend it. Can students defend their position? The group can set up a chatbot to argue for contradictory positions from a specific area of expertise, thus using adversarial AI to conduct mock debates. Can students argue their point convincingly? What steps can be meaningful to mitigate, control, or more or less solve the problem?
Whereas the assignment is about a wicked problem, the exercise is designed to help students learn to use AI in collaborating, presenting, and researching. It is in the context of the wicked problem that students learn to think with AI.
7. Use AI to enhance the delivery in terms of argumentation, flow, and aesthetics
This last step is about learning to use AI to enhance the delivery of the final outcome. Students can use AI to create infographics, social media posts, blog posts, press releases, etc. Focus on helping students learn how to use AI with the delivery of the position in simple language. Assessments can include issues of how convincing the arguments are and how useful the images are. The purpose of this step is polishing their position while keeping a documentation trail of how students used AI precisely.
N.B.: This project was the result of the ONL241 team 6 facilitated by Diane Pilkinton-Pihko and David Bevington. The course was designed as a collaborative learning effort of the participants, including Andrea Korens, Farhat Taj, Marko Scholze, Verica Aleksic Sabo, Saad Abdullah, and Carlos Diaz Ruiz.
