Agile Collaboration in the Age of AI
Since Agile practice began, the idea of increasing the level of collaboration within and across teams has been a cornerstone of how we tap the human potential. Now, it’s 2023, and there is a new team member in town. They are a member of every team and can instantiate themselves as many times as needed. “They” are AI chatbots.
Chatbots have been around for a long time. The best of them could barely hold down a coherent, superficial conversation. Over time, the word choices improved, but the topics were still limited. This year marked the proliferation of modern chatbots capable of tapping the vast knowledgebase of the internet.
Suddenly, these chatbots are smarter than we are. Instead of inane chatter about the weather, we can get an explanation of E=mc2 or a Python code snippet that solves a particularly sticky problem. Team members now have a choice to collaborate with humans on the team or their ever-present AI teammates.
I am reminded of the time when smartphones were first becoming ubiquitous. There was lots of discussion about when it was appropriate to have your phone out. Based on the old rules of etiquette, it was not appropriate to have it out in the company of others. Today, some would consider this to be a perfectly normal scene.
Was this image AI-generated? I don’t know. I suspect it was. When we need information now, we just ask a chatbot for it, and it is given. We are still asked to be skeptical about what an AI tells us, but this is a passing moment and not worth using as a reason to avoid AI.
The million-dollar question becomes when do we collaborate with AI, and when (and how) do we collaborate with humans? We can expect multi-modal collaborations to become the new normal. My fear is that we could easily find the humans on our teams decreasingly necessary. Teams of one human and many AIs may become the new normal.
If the image above is any indication, the people entering the workforce may be a lot more comfortable with any connection being made through the internet. It is entirely possible that these four people are actually chatting with each other and three other people who aren’t co-located. This apparent isolation may actually be inclusivity in disguise. As we move increasingly towards remote work, choosing our collaborator type is no longer a foregone conclusion.
It is reasonable to expect that organizations and teams will need protocols and guidelines around choosing collaborator types. Currently, most available AI tools lack sufficient context to fully participate in all aspects of team operations. That could change as organizations feed more of their data into training models. Company-specific AI systems will be a game-changing way to enhance AI’s ability to bring the right context to its contributions.
The benefits of human collaborators will narrow as AI capabilities expand. My recent article about innovation describes how diversity of ideas can contribute to creativity. AI is not likely to be able to make novel connections like humans do for some time. Collaborations where creativity is important will remain fertile ground for humans. Other, less advantageous collaborations with humans may need to be mandated.
Mandating that humans collaborate seems like an authoritarian anti-pattern. The time is approaching quickly when we must be intentional about how we utilize humans. If we’re not thoughtful, we’ll all become glorified AI prompt engineers…until they don’t need us for that either.