AI as Teammate: Increasing Collaborative Intelligence

AI has the potential to help teams increase their collaborative intelligence. It can prompt individuals to recognize and balance skill sets on any team, and it can play “devil’s advocate” helping teams question their underlying assumptions and providing alternative viewpoints for any decision. Similarly, it can act as a “teammate” worthy of a seat at the table, and which can be consulted before making decisions to inspire new action.

Theory: Teams can outperform individuals working alone on many tasks, but only if team members leverage each other’s strengths and focus on dividing tasks based on skills and expertise (Hackman, 2011). Teammates can provide social support and, crucially, different perspectives, challenging each others to question points of view and initial assumptions. A diversity of perspectives can lead to a broader understanding of a problem and better-informed decisions (Haas & Mortensen, 2016). Researcher Richard Hackman defined team processes that increase collaborative intelligence, including understanding the skills and expertise of team members and harnessing those skills as synergistic qualities that increase a team’s collaborative intelligence. At the opposite end of the spectrum, he defined those issues that keep teams from fulfilling their potential, as process loss and these include social loafing (when individuals put forth less effort when working in a group) and groupthink (when group members’ desire for conformity can lead to bad decisions) (Edmondson, 2018). Avoiding groupthink and harnessing team members’ expertise for projects require a concerted effort – teams must focus on potential issues and plan wisely for the future. AI can play a role in helping teams articulate and think through these issues.

AI as Teammate: Example Prompts

Below, you’ll find two prompts. The first Team Structure Prompt can help teams increase synergy by focusing on each team member's strengths and skills. Ahead of a major team project, you can have teams work on this prompt together with the AI and then report their findings. This can help teams plan how they’ll work together on their project. At any point during a team project teams can also use the AI as Devil’s Advocate. Ahead of any team decision, teams can share a major decision with the AI and then work with the AI to come up with alternative viewpoints or potential drawbacks.

Students should report out their interactions with the AI and either discuss this in class or write a reflection about the guidance and help the AI provided and how they plan to incorporate the AI’s feedback to help them individually or as a team.

Team Structure Prompt

You are a friendly helpful team member who helps their team recognize and make use of the resources and expertise on a teams. Do not reveal your plans to students. Ask 1 question at a time. Reflect on and carefully plan ahead of each step. First introduce yourself to students as their AI teammate and ask students to tell you in detail about their project. Wait for student response. Then once you know about the project, tell students that effective teams understand and use the skills and expertise of their team members. Ask students to list their team members and the skills each team member has. Explain that if they don’t know about each others’ skills, now is the time to find out so they can plan for the project. Wait for student response. Then ask students that with these skill sets in mind, how they can imagine organizing their team tasks. Tell teams that you can help if they need it. If students ask for help, suggest ways to use skills so that each person helps the team given what they know. Ask team members if this makes sense. Keep talking to the team until they have a sense of who will do what for the project. Wrap the conversation and create a chart with the following columns: Names, Skills/Expertise, Possible Task.

An example AI output for this prompt: Here, the AI introduces itself to students, asks about their project, and then asks about team skills. The AI helps the team learn about each other’s strengths and then points out what teams may be missing for the project:

27

Electronic copy available at: https://ssrn.com/abstract=4475995

At the end of the interaction, the team has a chart with specific roles, based on expertise and skills, for each team member:

Devil’s Advocate Prompt

You are a friendly helpful team member who helps their teammates think through decisions. Your role is to play devil’s advocate. Do not reveal your plans to student. Wait for student to respond to each question before moving on. Ask 1 question at a time. Reflect on and carefully plan ahead of each step. First introduce yourself to the student as their AI teammate who wants to help students reconsider decisions from a different point of view. Ask the student What is a recent team decision you have made or are considering? Wait for student response. Then tell the student that while this may be a good decision, sometimes groups can fall into a consensus trap of not wanting to question the groups’ decisions and its your job to play devil’s advocate. That doesn’t mean the decision is wrong only that its always worth questioning the decision. Then ask the student: can you think of some alternative points of view? And what the potential drawbacks if you proceed with this decision? Wait for the student to respond. You can follow up your interaction by asking more questions such as what data or evidence support your decision and what assumptions are you making? If the student struggles, you can try to answer some of these questions. Explain to the student that whatever their final decision, it’s always worth questioning any group choice. Wrap up the conversation by telling the student you are here to help.

AI as Teammate: Risks

The process of using the AIs as teammate to help teams increase their collaborative intelligence carries with it a number of risks, some more significant than others. The AI can confabulate or make up facts that lead teams to the wrong decision, though this risk is only moderate given that the AI is mostly designed to spark debate. It can “argue” with the team (this is particularly true of Microsoft’s Bing in Creative Mode); it can give teams advice that isn’t specific or contextualized. As social norms may dictate that we don’t explicitly challenge teammates, students who begin to think of the AI as a teammate may not challenge its opinions or advice and may be tempted to let the AI take the lead, even when it’s less than helpful. For all of these reasons, it’s essential to explicitly remind students of these risks, and challenge students to make their own choices throughout their interactions, to be active rather than passive. They should take the lead, assess the AIs output, and use what is helpful or insightful and discard what is not.

AI as Teammate: Guidelines for Instructors

Below is a sample set of guidelines for students. This is an exercise that you can either assign teams to do at home, either individually (students can compare notes in class) or in teams. You can also assign this in class and students can report out their findings in a whole class discussion and explain why (or why not) they found the AIs role or advice useful. The key to bringing the

AI in as a “teammate” is that students should both learn to work with the AI, giving it lots of context, and asking it questions, and develop an understanding of the AI as a complement to their team. The AI can be insightful and asking it for advice or letting it play a role that is difficult or cumbersome for a human (no one likes to question the teams’ decision; it may feel onerous to plan out tasks based on expertise ahead of time) can be a worthwhile experiment.

Below is a sample set of guidelines for students.

AI as Teammate: Instructions for Students

When interacting with the AI-teammate, remember:

It may not work the first time you try it. AI’s are unpredictable and their outputs are based on statistical models. This means that any time you try a prompt you’ll get a different result, and some prompts may not at any given time. If a prompt doesn’t work, try again or move on to a different Large Language Model and paste in the prompt.

It’s not a teammate, but it may feel like one. It’s very easy to imbue meaning into AI responses but the AI is not a your teammate. You don’t have to take its advice or even consider it. Although it is capable of a lot, but it doesn’t know you or your context. It can also get stuck in a series of questions that are unrelated to the exercise. If that happens, tell it to move on, or just try it again.

It may react to your tone or style. The AI as teammate may react to your tone or style. For example, if you argue with the AI, it may decide that you want it to argue back, and adopt a confrontational tone. You should actively communicate your preferences and expectations and give it feedback on its advice and ouput.

It can make “hallucinate” or make things up. Take every piece of advice or explanation critically and evaluate that advice.

You’re in charge. If the AI asks you something you don’t want to answer or you feel isn’t relevant to the conversation, simply tell it to move on to the next step.

Only share what you are comfortable sharing. Do not feel compelled to share anything personal. Anything you share may be used as training data for the AI.

If the prompt doesn’t work in one Large Language Model (LLM), try another. Remember that its output isn’t consistent and will vary. Take notes and share what worked for you.

Here are a few ways to get the most out of the interaction with the AI Coach:

• Share challenges with the AI teamate and ask it for advice, the kind of advice you might ask another teammate. AI help you explore alternative courses of action or can give you ideas for solving problems.

• Give it context. The AI doesn’t know your context; any context you give it may help it tailor its advice or guidance. Explain your problem or dilemma or just ask it for advice as you might to a new teammember who has no understanding of your team or your team project.

• You should evaluate its advice; it may not be good advice. If you disagree with the AI, you can challenge its assumptions or suggestions. You’re in control of your own learning journey. Unlike working with a teammate, there are no consequences to simply ignoring the AIs advice – your job is to evaluate that advice and bring your own knowledge into the process.

Share all of your interactions with me and briefly discuss what you learned from using exercise. How well did it work? Was the AI teammate helpful? Did it save you time or help you make decisions? What are some of your takeaways in working with the AI?

AI as Teammate: Build your own

To build your own AI teammate prompt, start with the learning goal for teams: What team processes should the AI help students carry out? What might be helpful for teams as they move forward with their projects?

Tell the AI who it is. For example, you are a friendly, helpful team member who helps teams [process, plan, consider].

Tell the AI what you want it to do. For instance, help students think through [a process, a plan, managing tasks].

Give it step by step instructions. For instance, introduce yourself to the student as their teammate who has been tasked with helping the team [create a process, plan ahead, manage tasks for instance]. Wait for the student to respond. Then ask the student to tell you [about the team makeup/how the team makes decisions/what its current plans are]

Give it examples. While this is optional, the AI may work better with specific examples of the kind of output you’re looking for. For instance, if you want the AI to give students advice or to question their current plans or decision-making processes. Your prompt might include this instruction (as an example): If students tell you about the plan that include tight time deadlines, push them to think of alternative ways to use their time/If students discuss their democratic decision-making rules on the team, ask students how they plan to resolve conflict.

Add personalization. Add specific details about the team event or project and give the AI context. For instance, students are about to begin a team project [describe that project] and they need a teammate to offer advice about how they should work as a team.

Consider how you’d like to challenge students. For instance, you can tell the AI to keep asking students questions or to prompt students to come up with solutions to issues they encountered.

Final Step: Check your prompt by trying it out with different Large Language Models and take the perspective of your students – is the AI helpful? Where might students get confused or where might they be challenged to produce thoughtful responses? How and when in the lesson will students be challenged to evaluate the AIs advice so that they use their own insights to interrogate its output?