CodeFreeIQ
← All cheat sheets
Cheat sheet · 4 min read

Before you trust AI with your data

You don't need to be paranoid. You do need to check three things before you paste sensitive data into any AI.

Abi Odedeyi By Abi Odedeyi · CodeFreeIQ

AI is safe to use on real work, if you're deliberate about it. The trouble starts when people paste client details, contracts or personal data into whatever tool is open, without a second thought. Three checks keep you on the right side of both good sense and your obligations.

What's inside

1Is your data training the model? 2Send only what's needed 3Know your own rules
1

1: Does this tool train on what I type?

Some tools and tiers use your inputs to improve their models; some explicitly don't, especially business and team plans. Before you trust a tool with anything sensitive, find that setting or policy and know the answer. It's usually a toggle or a plan choice, not a mystery.

What this means for you: 'free' sometimes means you're paying with your data. For client work, choose the tier that keeps your inputs yours.
2

2: Does the AI even need the real details?

Most of the time, the AI can do the job without the actual names, numbers and identifiers. Redact or swap them for placeholders and you get the same quality of help with none of the exposure. The best data-protection move is simply not sending the sensitive part.

What this means for you: if the AI doesn't need it to help you, don't give it. Redaction is faster than a data-breach conversation.
3

3: What are you actually allowed to share?

If you handle other people's data, you likely have obligations, a privacy policy, a client agreement, a regulation. You don't need to be a lawyer, but you should know your own rules before AI touches regulated data, and when in doubt, keep the sensitive part out.

What this means for you: this isn't about fear, it's about defaults. Set a sensible one now and you never have to think about it mid-task.