One of our new classes for California's CALL Academy is about examining AI with a focus on equity. We examine it through four main lenses: bias/discrimination, access and barriers, environmental impact, and workforce/labor impact.
This tool provides a simple checklist for considering these equity lenses when evaluating a new service, product, or tool. It's not binding or restrictive; it's just a guide to help inform your decision-making.
How to use: Work through each pillar with the tool in front of you. Tick a prompt once you've genuinely considered it — there are no automatic pass/fail thresholds. The “Flag for” line under each pillar surfaces the equity factors most likely to be affected. Finish with a verdict and the vendor questions at the bottom.
Your verdict
Where does this tool land for our community? This is a judgment call — the checklist informs it, it doesn't decide it.
Go
Concerns are minor or well-mitigated. Proceed with normal review.
Caution
Real equity gaps. Proceed only with safeguards, disclosure, or a plan.
Stop
Harm to a group outweighs benefit, or core questions can't be answered.
Questions to ask the vendor
- What data was this trained on, and who was — or wasn't — represented in it?
- What does full access cost a patron, including devices, connectivity, and reading level?
- Can you document accessibility conformance (WCAG) and supported languages?
- What's your stance on data retention, and can patron data be excluded from training?
- How were affected workers and communities consulted in the product's design?
