If you run a small business on the West Coast, you have probably wondered: is my information safe when using AI? It is a fair question. AI tools like ChatGPT, Copilot and Gemini are genuinely useful for drafting emails, summarising documents and answering customer queries. But the moment you paste something into one, it is natural to ask where that information goes and who might see it.
The honest answer is: it depends on which tool you use, which plan you are on, and what you type in. The good news is that once you understand a few basics, you can use AI confidently without putting your business or your customers at risk. Here is what you actually need to know.
What really happens to the data you type into AI?
When you type a message into an AI tool, that text is sent over the internet to the company running the tool (OpenAI, Microsoft, Google, Anthropic and so on). Their computers read it, work out a response, and send it back. So far, so normal — it works much like sending an email or using online banking.
The important part is what happens next. Two things can happen to your text:
- It gets stored for a while. Most providers keep a copy of your conversations for a period, partly to run the service and partly for safety and abuse checks.
- It might be used to train future AI models. This is the bit that worries people. On some plans, the words you type can be fed back in to help improve the AI — which means fragments of what you wrote could, in theory, surface elsewhere down the track.
Whether that second point applies to you comes down almost entirely to the type of account you are using.
Consumer vs business AI: the difference that matters
This is the single most important thing to understand, so it is worth being clear about.
Free and personal (consumer) plans. On free or personal versions of tools like ChatGPT, your conversations may be used to help train the AI by default. You can usually turn this off in the settings (look for “Data Controls” or similar), but many people never do, because they do not realise it is on.
Business, team and enterprise plans. On paid business tiers — such as ChatGPT Team or Enterprise, Microsoft Copilot for business, and the paid API services — the providers generally state that they do not use your data to train their models by default. Your information is treated as your business’s property under a proper commercial agreement.
In plain terms: if you are using AI for real work involving customer or business information, a business-grade plan is far safer than a free personal login. Policies do change, though, so it pays to have someone check the current terms rather than assume. That is the kind of thing we help local businesses sort out at Lighthouse IT.
Do’s and don’ts for you and your staff
You do not need a computer science degree to use AI safely. A few sensible habits cover most of the risk.
Do:
- Use a business or paid plan for anything work-related.
- Treat AI like a helpful stranger — useful, but not someone you would hand your customer database to.
- Write general prompts where you can (e.g. “draft a polite overdue-invoice reminder” rather than pasting the real invoice with names and amounts).
- Check the AI’s output before you send or publish it. It can be confidently wrong.
Don’t:
- Paste customer personal details (names, addresses, phone numbers, emails) into a free consumer tool.
- Enter financial data, bank details, IRD numbers or anything you would not put on a postcard.
- Share passwords, logins or access codes — ever.
- Assume the answer is correct or up to date without checking.
If your team uses AI, a one-page “rules of the road” document works wonders. Everyone knows what is fine to paste and what is off-limits.
Setting sensible guardrails for your business
You do not need an expensive system to use AI responsibly. Start small and practical:
- Decide which tools are approved. Pick one or two AI tools and stick to them, rather than letting everyone use whatever they stumble across.
- Choose the right plan. Move anything involving customer or business data onto a business-grade account.
- Write simple staff guidelines. A short list of “safe to use” and “never paste this” examples is enough for most small teams.
- Review the settings. Turn off model training where the option exists, and check data-retention settings.
- Keep a human in the loop. AI drafts; a person decides. That single rule prevents most mistakes.
AI and the New Zealand Privacy Act 2020
If your business handles personal information about customers or staff — and almost every business does — the Privacy Act 2020 applies to you, AI or not.
The Office of the Privacy Commissioner has published guidance for organisations using AI tools. The key message is common sense: only collect and use personal information for a genuine reason, keep it secure, and be transparent about what you are doing. Before adopting an AI tool for anything involving personal data, the Commissioner recommends thinking through the privacy risks first rather than after.
In practice, that means asking: do I actually need to put this person’s details into the tool, or can I keep it general? If a customer’s information ends up in an AI system, you are still responsible for it under New Zealand law. That is not a reason to avoid AI — it is a reason to use it thoughtfully.
Getting AI working safely for your business
Used sensibly, AI is a brilliant tool for small West Coast businesses — saving time on writing, admin and everyday tasks. The trick is knowing the difference between the free tool you play with at home and a properly set-up business account you can trust with real work. If you’re only just starting out, our guide on how to start using AI in your business is a friendly first step, and which AI should I use for my business? will help you pick the right tool.
If you are not sure whether your information is safe when using AI, or you would like help choosing the right tools, setting up business accounts and writing simple guidelines for your team, our AI consulting is here to help. Lighthouse IT is based right here near Westport, and we speak plain English, not jargon.
Get in touch with Lighthouse IT and we will help you use AI with confidence.