Is AI changing engagement across the CKAN ecosystem?

CKAN General
, ,
1

Last week, @rgradeck brought a super interesting article to my attention, about how AI may be changing the way people find and interact with open source platforms and their supporting communities.

The gist, essentially, is that in the pre-AI world, open-source communities relied on people asking questions in public as a primary means of identifying and engaging potential new contributors. This sort of this would happen via the traditional “doors” into their communities - through mailing lists, forums, GitHub issues, Stack Overflow, and the like.

Today, the argument goes, many of those same questions are being answered privately through AI assistants, disconnecting a critical point of contact between maintainers and community members. If this is true, it would follow that there would be implications for community engagement and project sustainability.

And it made me curious whether others in the CKAN ecosystem are seeing similar changes.

  • Are fewer people asking questions publicly than they did a few years ago?

  • Have AI tools changed how you learn about, troubleshoot, or contribute to CKAN?

  • If AI is making more of our users “invisible,” what new ways might we create for people to discover and participate in the CKAN community?

I’d be especially interested in hearing from maintainers, extension developers, data publishers, and organizations running CKAN instances. Are you seeing this shift in practice, or does it not match your experience?

2 Likes
2

It feels like without these interactions, it becomes much harder for leaders of the ecosystem to get important feedback from people that use the product. Curious if anyone here is feeling that feedback mechanisms have changed recently, and how processes of learning about user needs can adapt?

1 Like
3

I can confirm that these interactions have certainly changed for me. IMHO, it’s on the same unfortunate trajectory of technology having the side-effect of “disintermediating” us from each other.

In the name of convenience, “disintermediation” often means removing “friction” - the very thing that leads to interactions with another user.

ATMs removed the idle talk with the bank teller every two weeks; self-checkout - catching up with the same cashier asking about your kid; Wikipedia - killing encyclopedias; Search Engines - minimizing the need to go to a physical library to research a topic only to find yourself serendipitously finding another book on an unrelated, though relevant topic; Chatbots removing the need to go to StackOverflow - as it not only answers your technical question, it actually implements it for you. :frowning:

Much the same way social media was supposed to connect us, tech’s relentless advance has instead created personalized echo chambers for all of us. Personalization and gamification give us that small dopamine hit, as we “doom scroll” our curated, personalized feeds from YouTube, Instagram, LinkedIn, etc.

And it’s even worse with AI now gaining “memories” about you - your personal details and the topics you discussed before. With the upcoming new Siri - it even has access to your mail, text messages, todo-list, contacts, your media library - your entire computer. Some folks have even gone to the extreme of giving their OpenClaw “agents” their financial information to transact on their behalf.

IMHO, these are societal, wicked problems that we cannot hope to solve in our little corner of the world.

But perhaps, we can do our part… this forum itself is meta. It’s an attempt to aggregate all the conversations across all the GitHub repos across the CKAN ecosystem. It aggregates all the sites, tools and extensions in one place.

And maybe even in CKAN itself - FAIRifying all that data produces great metadata - which turns out to be the perfect context for AI. For FAIR data is AI-Ready data.

Despite our personal echo chambers - humans still care about where they live, their neighborhood; what’s happening and will happen around them; where their kids go to school, and what their taxes are paying for.

Maybe, if we curate our respective CKAN data portals so all that raw open data can be converted by AI into hyperlocal, personalized answers - a digital equivalent of the local paper, that sadly, has also been largely replaced by online feeds.

2 Likes
4

As it happens, the latest episode of Your Undivided Attention speaks directly to the topic -

5
  • Are fewer people asking questions publicly than they did a few years ago?

Yes, definitely. Sadly this is by design.

2 years ago, when LLMs where starting to take-off I had a conversation with an ex colleague of me that works with Microsoft. He told me: “In the future, organizations will write outputs for LLMs, not for humans.”. That was the sales pitch and vision of Microsoft, to replace human-to-human communication with human-ai-human interactions.

  • Have AI tools changed how you learn about, troubleshoot, or contribute to CKAN?

Quite a lot yes. I still do rely sometimes on CKAN Docs since they have become a reference material after the years. I know which section of the Docs contain the info I need, so it is easier to just go there than ask an LLM.

Pair-programming and troubleshooting has changed drastically since the arrival of AI.

  • If AI is making more of our users “invisible,” what new ways might we create for people to discover and participate in the CKAN community?

I think this is the tragedy of 98% of Open Source projects outside the big ones completely sponsored and governed by companies and Big Tech. Communities are infrastructure and without proper funding and caring communities stagnate.

Participating in a community requires effort and money. So a good question would be: which incentives are there for people to engage in a community?

Maintaining a community requires effort and money: Governance, events, engagement, infrastructure, social media, trademarks, marketing, etc. So a good question would be: which incentives are for sponsoring a community?

Whenever I talked to people, I always mention that CKAN has 2 open-for-everyone meetings every week. Not many people knows about it. I try to have a more marketing speech, something on the lines of: “Core developers are there, you can ask questions about CKAN, see it as free consulting.” (although I know it is quite the same). But without incentives, why would someone invest time/money in participating?

Last, I think it is important to have clear expectations and goals when we talk about communities. CKAN is a niche software (compared with React, Django, Go, Python, etc) so when we talk about a more active community (discover and participate) what do we mean with that? What success looks like? For me, personally, is sad to not have any Government in the community. I would love to have at least 2 Governments actively contributing and designing the future of CKAN.

Those are my two cents!