National Network Discusses Federal Data, AI, and Data Capacity Building

This month, the annual convening of the National Neighborhood Indicators Partnership (NNIP) was held in Charlotte, North Carolina. At such an unprecedented time for data, evidence, and local capacity, Hawai‘i Data Collaborative’s Emma Kern and Nick Redding attended the convening to learn from national peers about the ways in which data support organizations are adapting to the rapidly shifting data landscape. Among the many discussions, three topics stood out: advances in artificial intelligence (AI) tools, the potential loss of federal statistical data, and approaches to building community data capacities.

The discussion on AI was broad, with members’ orientations to the tools ranging from enthusiastic to very concerned – some were against usage of AI tools altogether. What was clear is the thoughtfulness that all members brought to the discussion. Whether embracing it, or being outright opposed to it, members seemed to agree that AI is a significant development for those who work with data. Potentially positive developments from AI that were mentioned included new possibilities for coding unstructured data, and efficiency gains from AI assistance across analysis tasks. One significant downside discussed was that AI may make conducting research and analysis too easy, leading users to abandon the rigor typical of traditional, professional data work. Other concerns including the limited functionality of large language models, the privacy implications of entrusting a few large companies with sensitive information, and the environmental and social impacts of rapidly expanding AI infrastructure.

Another topic explored during the convening was the status of federal statistical data. We at HDC have previously posted about the unprecedented threats posed to federal statistical agencies under the current administration. For many attendees, public federal data is foundational to much of the work they do to support their communities. Generally, there was agreement that the future availability and trustworthiness of federal data is a concern, but it is also too early to assess the full impacts. Much of the conversation centered on developing alternative data sources and championing local capacities. Many ideas were raised, from advocating for alternative State and local government data sources, to developing local surveys and crowd-sourced data sources, to leaning more on private data sources such as marketing and credit bureau data.

Lastly, there were enriching discussions about models for building local data capacities. Similar to our history at HDC, many network partners have been grappling with the challenge of building robust data tools and dashboards while struggling to garner engagement – lack of data capacity being cited as the primary inhibitor. Attendees shared a variety of approaches that are being tested across the country, from nonprofit data maturity frameworks, to “Data Days” that convene neighborhoods around data for advocacy, to models for community data trusts. Discussions across the room raised shared themes: data capacity building takes time, resources, and trust, there is hardly ever a “one-size-fits-all” approach, and building data capacities is a strategy for unlocking local sources of data in the face of federal uncertainty.

Overall, it was an enriching discussion and encouraging to see this network innovating data solutions amidst such rapid political and technological change.