The environmental concern is real. Training large AI models uses significant energy and water. Running them at scale has a measurable carbon footprint. For organisations that care about sustainability, this isn’t trivial. The counterpoint: the energy cost of a single AI query is roughly equivalent to a Google search, and falling. If you’re comfortable using email and cloud storage, the incremental impact of AI tools is modest.
Bias is real too. AI models reflect the data they were trained on, which means they can reproduce societal biases around race, gender, disability, and more. In practice, this shows up as assumptions in generated text — defaulting to certain demographics, reinforcing stereotypes, or missing cultural nuance. The fix is human review. Treat AI output as a draft, not a final product, and you’ll catch most of these issues.
The authenticity question. Should a nonprofit use AI to write donor communications? We think so — as long as the message is genuinely yours. Using AI to draft an appeal letter that your team then shapes, refines, and signs off on is no different from using a copywriter. Using AI to generate and send communications without anyone reading them first is a problem. The line is human oversight, not human typing.
The creative question deserves special respect. This is particularly important for arts and culture organisations — and we work with many of them. AI models were trained on vast amounts of creative work without the consent of the people who made it. For organisations whose mission is to support and champion original creativity, that’s not an abstract concern. It’s personal. If your organisation decides that using generative AI for creative output conflicts with your values, that’s a legitimate and principled position. There’s a meaningful difference between using AI to draft an internal grant report and using it to generate artwork for your exhibition catalogue. You get to draw that line where it makes sense for your mission.
The broader ethical case. If AI lets a three-person comms team produce the output of a five-person team, that’s more impact per donor dollar. If it helps you personalise supporter communications in ways you couldn’t afford to do manually, that’s better relationships. The ethical risk of using AI carelessly is real. But so is the ethical cost of ignoring tools that could help you serve your community better.
