Professor Patrick Parrra Pennefather | Kris Krüg

<aside> <img src="notion://custom_emoji/3b7b79b0-95af-4500-931c-e5c63e5df242/133c6f79-9a33-809e-b97a-007a409367c7" alt="notion://custom_emoji/3b7b79b0-95af-4500-931c-e5c63e5df242/133c6f79-9a33-809e-b97a-007a409367c7" width="40px" /> “The rapid advancement of generative AI technologies has sparked a transformative wave across various sectors, but one of its most intriguing and still to be determined impacts is in the realm of community building.”

</aside>

Building a Grass Roots AI Community of Practice: A Vancouver-Centered Use Case

Dr. Patrick Parra Pennefather

With Kris Krüg

“The rapid advancement of generative AI technologies has sparked a transformative wave across various sectors, but one of its most intriguing and still to be determined impacts is in the realm of community building.”

The lack of a citation is due to an overall sense of confusion when it comes to citing a large language model. I’d rather give credit to the entire human populace that contributed to its corpus than to Gemini, ChatGPT, Llama3, Mistral or Claude. Citing all humans who have ever created work on the internet would make for a long read though. Nevertheless, it is still important to Introduce the quote as coming from a large language model (LLM), because what a machine learning model doesn’t know, is that it is too early to make assertions more widely than perhaps one single community of practice in its infancy. Despite other ‘hallucinations’ that might accompany the generation of puffery from any LLM, it may in fact be true that there is an in-person community of practice that is growing in Vancouver centered around use cases, debates, demos, and new innovations, that in some way or other has been informed by generative AI and AI systems. What can we learn from the growth of that community? Might we find useful research from other tech communities we can draw from in this area, or is this relatively new territory? More importantly, what key ingredients will create a more sustained community of practice around AI and why does it matter? This paper relies on a Vancouver-centric AI community as a use case to uncover the essential ingredients, that contribute to building a successful community of practice within the tech community. The ingredients might inspire those who are considering the development of their own localized communities centered around the curiosity, integration, experimentation and resistance to public and private generative AI.

You cannot step into the same river twice

Bias, generalizations, and alcohol-infused nostalgia aside, this essay, this report, this summary of experience of an emerging community of practice can only rely on making warranted assertions. Heraclitus might be one of the earliest sources of ideas that were similar. He emphasized the constant change and flux in the world. His famous dictum, "You cannot step into the same river twice," plays with the idea that observations and understanding are inherently context-specific. Each moment and situation is unique, which implies that assertions about phenomena must consider their specific contexts. Not that recently (1938), the notion of warranted assertions, as proposed by John Dewey and others in pragmatist philosophy, emphasize that claims or assertions are warranted when they are grounded in systematic inquiry. Dewey argued that knowledge emerges through a process of inquiry where hypotheses are tested against evidence from experience, aiming to resolve doubts and achieve warranted assertability rather than absolute certainty[1]. Austin introduces the concept of speech acts, distinguishing between locutionary, illocutionary, and perlocutionary acts. Assertions fall under illocutionary acts, where the speaker commits to the truth of the statement[2].

Building on Dewey’s ideas, qualitative researchers emerged from the shadows of empiricism with a similar conceptual framework of naturalistic generalizations[3]. The cleverly titled chapter “The only generalization is: There is no generalization”, acknowledges the contextual and situated nature of qualitative inquiry, where generalizations are not universal laws but emerge from detailed, context-bound observations. Naturalistic generalizations in qualitative research emphasize the importance of rich descriptions, thick data, and interpretive insights to construct plausible and contextually specific claims about social phenomena. Extending from the tradition of situated research, a hypothesis statement can be co-constructed with the support of an LLM.

LLM: Regular in-person gatherings centered around generative AI topics will strengthen interdisciplinary bonds among participants, leading to enhanced knowledge sharing, collaborative project initiation, and cross-disciplinary skill development within the community of practice.

While the more scientific-learning version above required careful cultivation and refinement over at least two dozen versions and specific prompting it might be better to create a hypothesis more akin to naturalistic generalizations we might make from the results of our inquiry.

LLM: The integration of generative AI technologies may catalyze the formation of an interdisciplinary community of practice, fostering cross-disciplinary interactions and collaborations among researchers, developers, creatives, and individuals with varying levels of familiarity with AI technologies.

Agent provocateur

We can all agree on one naturalistic generalization with the recent public generative AI pandemic; it is successful in disquieting how we think about creativity, work and knowledge in different ways. By ‘us’, I mean those humans who actually want to tell the story of how they are reconciling its use. Public generative AIs like many an LLM are glorious and sketchy, useful, and at times their use irreconcilable, probabilistic computing machines that tend to exclude voices outside the normative; biased spouting machine learning models we are prone to anthropormizing, for better or worse. However, the justified hullabaloos around different types of generative AI, the technology is sparking an AI community to life in Vancouver; a cross-disciplinary group of people that seem committed to determining the value of gen AI, weighing its pros and cons, and understanding where other diverse members of that community place themselves somewhere between the spectrum of adoption and rejection of the technology. While it's still too early to make promethean generalizations, there are emerging signs that generative AI may be provoking new forms of community engagement and collaboration.

That engagement is evident in the activities of the Vancouver AI community meetup. Monthly meetups since January 2024, organized by Kris Krüg and the Future Proof Creatives initiative, highlight an emerging community exploring the impact and potential of AI. These events feature a range of activities from expert talks and hands-on workshops to showcasing emerging experiments with generative AI, and networking opportunities, reflecting the diverse ways in which generative AI is being integrated into business and creative processes.

Precedence and Antecedence

Communities of practice around the development of machine learning have been establishing themselves for decades, particularly at international conferences: