Community is undervalued in data and tech but we’ve never needed it more

Digest #76

I run a community garden in Harlem, New York. We keep some boisterous chickens and run outreach programs for local kids. What we plant is crowd-sourced and where it’s planted is based on sun mapping data and what evidence shows will work where. We sometimes squabble about small things, yet despite our differences, we share a belief that growing and eating fresh food should be accessible to all. 

I am proud of our urban farming community. By contrast, I was really perturbed when I learned that another community in which I am active, Reddit, has just sold its entire content base to Open AI for use in ChatGPT. Several decades of community contributed and moderated content have created a cash cow, not least for Open AI’s CEO, Sam Altman, whose wealth already includes $615 million in Reddit stocks. 

We all laughed when Google AI told people to put craft glue in pizza after regurgitating a sarcastic comment. But as a member of a subreddit where people share experiences of their life-threatening allergies, I know that unverified medical advice from redditors, stripped of context, could prove lethal. 

This is just one example of those who are developing emerging technologies exploiting the goodwill and societal benefits of communities. Unsurprisingly, a vacuum in regulation around generative and other types of AI is leading to declining public trust in AI systems and the companies that build them. 

Community that puts people first should be protected 

The Data Values Project reminds us that we shouldn’t accept a status quo that’s been established with profit, not people in mind. It also shows us that we can protect our rights through community action. 

In the past, social media companies mostly used our data to sell us things. Today, our family photos are morphed into AI generated images, while even the most innocuous photos can expose young people to becoming victims of deepfake nudes and our texts and posts are now fodder for chatbots to share with no attribution. In this new landscape, the value proposition for human engagement in online communities is arguably declining, despite the technological potential. 

A colleague recently said in an interview that we need to build a better digital ecosystem, so that we can have a thriving, creative, safe society. A lot of my work is concerned with examining the fundamentals of the relationship between data and society and I think developing and protecting community, particularly the data for development community, is essential in building this ecosystem and ensuring society benefits from technological and digital developments. 

The data for development community has come a long way, yet faces new challenges

When the Sustainable Development Goals were introduced in 2015, the World that Counts report made recommendations about integrating non-traditional data sources while protecting data quality through the institutions of national statistical offices. Ultimately the data community needed to be strengthened and brought together, but at the time there were a number of mini-communities who were at odds with each other. Disagreements broke out between the open data evangelists and statisticians about how much data could and should be shared, or about the tradeoffs between timeliness versus ensuring accuracy and about the role of non-data specialists in generating data sets and whether these could be included in official statistics. 

But despite our differences, we were united by the common belief that better data offers an important pathway to progress and is of central importance in achieving the Sustainable Development Goals. Over the past 9 years there has been more cooperation, less suspicion and slow but steady integration of new data sources such as  Citizen-Generated Data or earth observation data. Our data for development community has grown and strengthened since 2015 and this has driven SDG progress for those most in need around the world. Now is a moment where we need to come together once again to address the abuses and potential for exploitation presented by AI and other emerging technologies. 

Community must once again be central to our progress

Communities are not static. The composition of the Global Partnership’s network has evolved over time, becoming more diverse, with a wider geographic spread and more campaigning organizations, recognizing that our shared challenges are not merely technical but also normative, ethical, even existential.

The network’s recent partner survey revealed that numerous network members are already engaged in AI for development, from many different angles. Many others, including many government agencies, are facing challenges with technical capacity, skills and investment that are hindering their ability to fully embrace the opportunities of AI, while navigating a complex regulatory landscape. Our community building can prove critical here  - as in some of the recent work on contextualizing AI for country and regional settings. The strength of the community can also be seen in conversations around data governance to protect rights around the world, and plans to deliver this through the Global Digital Compact.

These new topic areas risked the proliferation of sub-communities and silos again, but our community has hard-won experience building coalitions that leverage data and innovation for economic and social progress that help bind us rather than divide us. Global inequality is about to be supercharged by the disparities in the global data and AI landscape. It follows that we need to co-create a core community that encompasses the true breadth and potential of the data and digital revolution, rather than splintering into siloed micro-communities that talk only to themselves. Together we can push for human progress and fight inequality and the climate crisis, one data point at a time. 

Community has never been more important than at this particular moment in history. To build trust, coherence and a safer digital ecosystem, we must come out of our comfort zones and come together to build a fairer data future.