Trust comes from checking, not hoping. For a small organisation, a trustworthy AI tool handles information safely, gives answers people can rely on, behaves as intended, and has a named owner.

  • Handles information safely.
  • Gives answers people can rely on.
  • Behaves as intended.
  • Has a named owner.

AI is now part of everyday work for many businesses and charities. Teams use it to draft letters, answer enquiries, summarise notes and find information faster. The useful question is no longer “should we use AI?” but “how do we know it’s doing what we think it’s doing?”

What trustworthy AI means for a small organisation

It handles information safely. Personal and sensitive data goes only where you have agreed it can go, and stays under UK GDPR.

It gives answers people can rely on. Output is accurate enough for the job, and someone checks it where mistakes would matter.

It behaves as intended. It doesn’t leak internal information, follow hidden instructions in documents or emails, or drift into topics it shouldn’t touch.

If you cannot say how your tools meet each of these, you do not yet know whether to trust them.

Where AI risk usually hides

The risks we see most often are not dramatic. They are everyday gaps:

  • Staff pasting sensitive data into public tools. Client records, HR issues or safeguarding notes go into a free chatbot, and you have no idea where they’re stored or how they’re used. See can charities use ChatGPT with sensitive data.
  • Confident wrong answers. Large language models can produce plausible text that’s simply untrue. A customer-facing assistant that invents a policy or a price causes real problems.
  • Prompt injection. When an AI reads emails, web pages or uploaded files, those can contain hidden instructions that make it reveal information or act in ways you didn’t intend.
  • No owner. Nobody is responsible for the tool, so nobody notices when it starts going wrong.
  • Shadow AI. People use tools the organisation hasn’t approved because they’re quicker than asking.

Test it before you trust it

You wouldn’t launch a new website without clicking every link. AI deserves the same treatment. Before putting a tool in front of staff or customers:

  • Write a set of realistic questions, including awkward ones, and check the answers against what you know is true.
  • Try to make it misbehave: ask for information it shouldn’t share, feed it a document with odd instructions, ask questions outside its remit.
  • Record what went wrong, fix the setup (instructions, data sources, permissions) and test again.
  • Repeat on a schedule. Models and data change, so a tool that passed in the spring may not pass in the autumn.

Larger organisations do this formally as LLM security testing. The idea scales down well: a structured afternoon of testing is far better than none.

Put some simple governance around it

There is an international standard for managing AI responsibly, ISO/IEC 42001, which sets out how an organisation should run an AI management system. Few small businesses or charities need certification, and this is not a claim that Blue Door holds that certification. The thinking is still useful at any size: know which AI you use, assess the risks, assign owners, and review regularly.

In practice, for a small team, that looks like a one-page AI policy, a short list of approved tools, a named person accountable for each one, and a review date in the diary. The NCSC guidelines for secure AI system development are a useful official reference when you write that down.

Keep sensitive data under your control

For charities handling client records and SMEs holding customer or staff data, a large part of trust is choosing where the AI runs. Public consumer tools are convenient, but you often cannot control where data is processed. A private setup, on a system you control, with agreed processing, is much easier to explain to trustees, funders or customers when they ask where their information goes. More on that in private AI operations and local versus cloud AI models.

A quick trust checklist

  • List every AI tool your team uses, including free ones.
  • Decide what data may and may not go into each tool.
  • Name an owner for each tool.
  • Test each customer-facing or decision-supporting tool with real questions, including awkward ones.
  • Try to break it: hidden instructions, off-topic requests, requests for confidential data.
  • Require a human check wherever an error would affect a person or cost money.
  • Write a one-page AI policy and share it with staff.
  • Set a review date, at least every six months.

If you have already rolled something out and it is not working, the same checks are the start of AI Rescue: go back to the job, find the break, and decide whether to keep, fix, rebuild or stop.

Questions

Is it safe to use ChatGPT or similar tools with client data?

Not by default. Check the provider’s terms, your account settings and where data is processed. For sensitive or personal data, a private setup you control is the safer option.

Do we need ISO/IEC 42001 certification?

Most SMEs and charities don’t. The standard is a useful framework to borrow from, but a clear policy, named owners and regular testing cover the essentials for a small organisation.

How often should we test our AI tools?

Before launch, after any significant change to the tool or its data, and on a regular cycle, at least twice a year.

If you’re not sure whether your AI can be trusted, or you want to use AI on sensitive information more safely, book a free 40-minute call by phone or Teams.

Book a free call