🌿 The Spore

Brought to you by the team at Organizational Mycology

👀 Something that Caught Our Eye

Is Artificial Intelligence being “Openwashed”?

In the early periods of the environmental movement, and even today, companies saw opportunities to curry favor with the general public by portraying their products and services as “sustainable,” “environmentally-friendly,” or “green.” Yet many of these products and services did not live up to their advertising, a phenomenon that became known as “greenwashing,”  

A similar phenomenon has occurred in the Open Science space over the years. Publishers instituted confusing (and costly) tiers of “open access” publications, data-centered organizations touted the accessibility of some of their data, and software projects remained in many ways gated while advertising that source code is open. This related phenomenon became known as “openwashing,” a brief history of which you can read about on Wikipedia. 

Recently, companies founded to push the frontiers of artificial intelligence (AI) have reanimated the conversation around openwashing. A recent article in The New York Times (paywalled! the irony!) calls out industry leaders such as OpenAI and Meta for identifying its technologies as "open" when, in reality, there are substantial restrictions on and barriers to using the tools. Meredith Whitaker and colleagues took a more scholarly, critical approach and wrote that the practice of labeling AI systems as "open" is misleading, "often constituting more aspiration or marketing than technical descriptor." Moreover, the authors argue that the practice takes advantage of free open source labor and appropriates the ethos of the open source software movement. As a result of this trend, academics have analyzed the availability, documentation, and access of various AI models, and organizations like The Linux Foundation have developed frameworks to design and evaluate the openness of the models.

Beyond these analyses and frameworks, what can the open source software community do to bring these issues to light and make the use of "open" in AI more aligned with the community's values? Is it worth it? Are there examples of AI models that really do fit the definition of "open" as you define it?

What We’ve Been Up To: OrgMycology Facilitating CZI EOSS Community Calls

A couple of weeks ago, OrgMycology facilitated the first Chan-Zuckerberg Initiative community call for the Essential Open Source Software program. Our conversation centered on the management work required to run an open source project: the leadership, funding, delegation, prioritization, and roadmapping tasks that projects deal with across all phases of their evolution. 

These community calls are intended to be conversations where attendees can learn from one another, share experiences and challenges, and generally get to know one another. We’re also producing short outputs to disseminate in the broader OSS community. The first is a checklist of questions to consider when managing an open source project. 

As we cover more topics, we’ll collectively produce additional resources that we hope will be useful for leaders and members of open source projects. In our next call, for example, we’ll discuss “cat herding”: the challenges and strategies open source projects use to incentivize and organize people to contribute to, maintain, or otherwise participate in the project. While we’ll be specifically talking about open source projects, community members' insights will likely be useful for anyone who finds themselves wrangling people together toward common goals–especially in organizations and communities that depend heavily on volunteers. 

It got us thinking about “styles” of cat-herding and how many different leadership styles can bring groups together. We’re excited to hear what styles the EOSS community has seen in their projects and in the wild, from incentive-laden styles like offering recognition and status to grassroots styles that motivate via community-building and collective action. What cat-herding styles have you seen in your work, and where did they flourish or struggle? 

Short updates 

OM/The Turing Way/JupyterHub – OrgMyco is kicking off an engagement with The Turing Way and JupyterHub to better understand and address community needs in the JupyterHub community. We’ll be interviewing community members about their experiences before going into a more strategy-focused phase, and we’ll be sure to share things we learn that might be useful to other communities.

Beth will be at the XOXO Festival  in Portland August 22-24th – If you’re going to be at this one-of-a-kind festival celebrating arts, culture, gaming, tech and all things digital, send a note and Beth will gladly say hello.