by azadi sheridan
With a new AI tool launching almost every week in 2026, it is easy to get caught up in the "feature hype." However, for charities, non-profits, and responsible businesses, the best tool isn't just the one with the most "bells and whistles"—it’s the one that aligns with your values and legal obligations.
When evaluating a new piece of AI software for your team, use this three-pillar framework to ensure you’re making a sustainable and secure choice.
1. Data Storage and Privacy: "Where does my info go?"
The biggest risk with AI is the "black box" of data. If you feed sensitive donor information or internal strategy documents into a tool, you need to know exactly how that data is being handled.
Training Usage: Check the Terms of Service to see if the provider uses your inputs to train their "global model." For work purposes, you should ideally look for Enterprise or Team tiers that offer "Zero Data Retention" or explicitly state that your data is not used for training.
Encryption & Compliance: Is the data encrypted at rest and in transit? Does the provider comply with UK GDPR?
Data Residency: Some organisations require data to stay within the UK or EU. Check if the AI provider allows you to choose your server location.
2. Ethical Impact: "Is it fair and transparent?"
AI models are only as good as the data they were built on. If that data contains historical biases, the AI will likely replicate them.
Bias Mitigation: Ask the vendor how they test for bias. If you are using AI for recruitment (like screening CVs) or service delivery, an unbiased tool is a legal and moral necessity.
The "Human in the Loop": Does the tool allow for human oversight? You should never "set and forget" AI-driven decisions. Ensure the software provides explainable outputs—meaning it can show why it reached a certain conclusion.
Accountability: If the AI makes a mistake or "hallucinates" (provides false information), what is the process for correction?
3. Environmental Impact: "What is the hidden carbon cost?"
It’s easy to forget that AI "lives" in massive data centres that require enormous amounts of electricity and water to run. In 2026, a single complex AI prompt can consume as much energy as leaving a lightbulb on for an hour.
Infrastructure Efficiency: Research whether the AI provider uses "Green Data Centres" powered by renewable energy.
Model Size: Sometimes, a "smaller" model is better. If you only need a tool to summarise text, you don't necessarily need a massive, power-hungry multimodal model. Choosing the right tool for the specific task helps reduce unnecessary carbon emissions.
Sustainability Reports: Look for companies that are transparent about their water usage and carbon footprint.
Summary Checklist for Your Team - four key review areas & key question for each:
Data Is my data being used to train the public AI?
Security Does the tool meet UK GDPR and SOC2 standards?
Ethics Has the model been audited for racial or gender bias?
Planet Does the provider have a net-zero commitment?
