Nonprofits Governing Powerful Technologies Responsibly Part 3

Pillar 3 Accountability Through Governance

As mission-driven organizations, you strive to operate with integrity, transparency and accountability to your stakeholders and communities. But let’s be honest, when it comes to governing powerful new technologies like AI, achieving true accountability gets exponentially harder.

We’re talking about complex, opaque systems that can shape high-stakes decisions impacting people’s lives in very real ways. From allocating resources and services to amplifying narratives and more – AI’s potential impacts are massive.

That’s why it’s absolutely critical for nonprofits of any size to have robust governance structures in place to ensure accountability, oversight and the ability to course-correct AI tools exhibiting harmful behaviors or unintended negative effects.

I get it, you may be thinking “Sure, that’s all fine and good for large nonprofits with big budgets and dedicated teams. But what about us smaller shops already stretched for time and resources?”

That’s a valid concern. Implementing full-scale AI governance boards and auditing frameworks from scratch may seem daunting. But there are still pragmatic steps even lean nonprofits can take to uphold accountability:

  • Designate an AI governance lead or working group to own this
  • Adapt governance templates from expert orgs to your use cases
  • Establish clear processes for reviewing models and impacts
  • Communicate openly and invite external audits/feedback

A person holding a glowing ball

Description automatically generated

Image by Leo from Pixabay

Minimum Viable Product (MVP) Governance Model

The key is starting small with a Minimum Viable Product (MVP) governance model that provides checks and balances. Leverage free resources, templates and best practice guides from groups like the Charity Excellence Framework and others.

A few key points on how lean nonprofits can start small with an “MVP” governance model:

  • Designate an AI governance lead or working group

Even if it’s just 1-2 people initially, identify who will own and drive the AI governance efforts. The members of the committee could consist of an existing staff member, a board member, and/or a volunteer – or a mixture.

  • Adapt templates from expert organizations

Rather than building from scratch, leverage free templates, frameworks and best practice guides from groups like the Charity Excellence Framework, Nonprofit AI Use Policy Template by Freewill, Acceptable Use of AI Tools Template by Community IT and others. Customize these to fit your use case needs.

  • Define clear processes

Even if lightweight initially, document your processes for activities like reviewing proposed AI use cases, auditing models for bias/harms, evaluating impacts post-deployment, etc.

  • Communicate openly and invite external audits

Be fully transparent about your AI governance approach. Invite external audits from local experts and feedback from your supporters to validate you’re on the right track. Post your AI research timeline and process on your website to be open and honest with your supporters.

A close up of a rose

Description automatically generated

Image by Nicky ❤️🌿🐞🌿❤️ from Pixabay

Collect your dot…

The key is starting with a lean governance model that at minimum provides checks and balances, community representation, and processes for oversight.

From there, you can gradually develop your practices by establishing oversight roles, investing in more technical monitoring capabilities when affordable, and continually expanding your accountability mechanisms in lockstep with your AI adoption.

Because at the end of the day, accountability isn’t just a nice-to-have – it’s a fundamental obligation we have as trusted nonprofits developing these world-shaping technologies. You must find a way, regardless of your size, to be open and honest with your supporters by having a plan when incorporating AI into your processes.