AI Self-Governance: It’s a Continuous Process

Why It Matters

MySociety’s iterative governance model demonstrates how NGOs can responsibly scale AI, setting a benchmark for sector‑wide ethical deployment and unlocking new funding streams.

Key Takeaways

• Living AI framework ensures continuous policy relevance
• Internal AI register tracks all organizational AI deployments
• Regular meetings enable cross‑team risk assessment
• LLMs now power data collection and insight generation
• Funding trends favor responsible AI for democracy

Pulse Analysis

AI governance is no longer a one‑off checklist for nonprofits; it requires a dynamic, iterative process that adapts to rapid technological change. mySociety’s living AI Framework exemplifies this shift, providing staff with a reusable set of principles while allowing continuous updates as new tools emerge. By embedding the framework into daily operations—through an internal register that logs every AI use case and scheduled cross‑functional meetings—the organization creates real‑time visibility into risks, benefits, and compliance, a model that other civic‑tech groups can replicate.

The practical impact of this governance structure is evident in mySociety’s evolving workflows. Developers now experiment with coding‑agent tools, measuring whether they truly accelerate development or solve problems traditional code cannot. Meanwhile, large language models generate complex datasets, such as parliamentary group (APPG) scrapers and WriteToThem insights, unlocking analytical capabilities previously out of reach. On the front‑line, AI‑driven screening filters misrouted immigration requests and prototypes machine‑learning models to flag abusive messages, illustrating how responsible AI can improve service quality while safeguarding user trust.

Beyond internal benefits, the external environment is increasingly rewarding responsible AI adoption. Funding bodies are earmarking grants for projects that demonstrate ethical AI use in democratic processes, and international civic‑tech communities are sharing best practices through Communities of Practice and TICTeC networks. This convergence of governance, operational innovation, and market incentives signals a maturation of AI in the public‑interest sector, encouraging organizations to adopt structured, transparent frameworks that balance impact with accountability.