Over six weeks earlier this year, I taught AI literacy to a group of BME women at WODIN in Liverpool.
Most had never used ChatGPT before session one. By session six, they were teaching their families.
This is what I learned.
The women who use AI best are not the most technical
Glory had spent eight years caring for her elderly mother and raising two children. She had no formal CV. She did not see herself as employable.
By session three, she had used AI to turn that experience into a Care Assistant CV. The right language. The right structure. Submitted with confidence.
Her words afterwards: "Today I learned that my past experiences, volunteering and skills, even though not from a formal setting, can be tailored to generate a strong CV."
Mary's experience was different. She was nervous about AI. By session five, her words were: "This training is helping me to think, write and talk professionally now, because I see AI as a close friend I can reach out to easily."
Rosemary's takeaway was the one that mattered most. After session six, she wrote: "I learned the importance of being able to ask AI probing questions and not take information given by AI as perfect without questioning its merit ethically."
That last sentence is what good AI judgement looks like.
What we actually taught them
Not the technology. Not how transformers work. Not prompt engineering tricks.
We taught them two simple habits.
The first habit is what to do before you type anything into AI. We called this S.A.F.E.™ (Secure, Adapt, Fine-tune, Execute). Do not paste personal data. Set the context first. Give clear instructions. Then send.
The second habit is what to do after AI responds. We called this V.E.R.I.F.Y.™ (Validate, Examine, Reference, Interrogate, Freshness, Yield). Does this match what you already know? Where did it come from? Is it current? Is it safe to act on?
These are the same frameworks I now use with FCA-regulated firms.
You can download a one-page summary of both frameworks here: www.aimindset.co.uk
Why this matters
A compliance officer worried about a Copilot data leak and a single mother drafting a benefits appeal are facing the same underlying question:
"How do I use this tool safely?"
The technology does not care about your background. The risks do not either. Neither should the safety frameworks.
This is also why I do not believe AI literacy training should be only for corporate audiences. The people most likely to be harmed by an AI hallucination, whether that is wrong benefits advice, fake immigration rules, or voice-cloning scams targeting their elderly parents, are often the ones with the least access to good training.
That is the gap WODIN's programme started to close. It is the gap I want to keep closing.
Final thought
The women who use AI best are not the fastest. They are not the most technical.
They are the ones who pause, question, and check before acting.
If you run a community organisation, housing association, or council programme where people are using AI without guardrails, I would love to talk.
If you are inside a regulated firm wondering how to deploy AI without putting your staff or your clients at risk, the same conversation applies.
The frameworks travel.
If this was useful, subscribe to The AI Mindset for practical guidance on AI safety, governance, and real-world use in plain English.
Stephen From AI Mindset