This blog is brought to you by Ashley Alexander, Chief Marketing Officer at Girlilla Marketing.
‍

‍
‍I know AI is everywhere right now, and I know it can feel a little overwhelming. Most of us are already using it at work in some capacity, whether that’s brainstorming ideas, polishing an email, summarizing a meeting, analyzing data, writing social copy, or just getting through some of the everyday tasks that take up more time than they should. When it’s handled well, it can sharpen how we work, and sometimes even make the work more fun.

But the easier it becomes to use, the easier it is to stop thinking about what we’re actually putting into it.

And that’s where we need to be a little more careful.

What doesn’t get said often enough is the simplest question: what information are we providing these tools?

Every time we upload a file, paste an email chain, share meeting notes, or provide background information, we’re sharing data with another platform. And as AI becomes second nature in our everyday workflows, it’s easy to forget to stop and think about what’s actually contained in that information and how sensitive that information might be.

That’s the mindset we need to bring to every interaction with AI: What does this tool actually need from me to accomplish the task?

Before you hit upload, paste, or submit, it helps to run a quick check:

• Is there personal information here?
• Is this confidential company material?
• Does it include client or customer details?
• Am I actually allowed to share this with a third-party tool?
• Do I understand how this specific AI product stores, uses, or keeps my data?

The point isn’t to swear off AI. The point is to treat it with the same care you’d apply to any other service you trust with information.

So what does responsible AI use look like?

Being cautious doesn’t mean staying away. It means building a few practical routines into how you use these tools day to day.

One easy rule: treat an AI tool like a person outside your organization. If you wouldn’t send the info to an external contact by email, don’t casually drop it into a chatbot.

1. Think before you paste.

Before you toss in an email, document, transcript, spreadsheet, or slide deck, take a moment to scan what’s inside. Confidential details? Client data? Personal info? Passwords or login credentials? Financials? HR or employee records? If any of that shows up, slow down and reconsider.

I’ve had to train myself on this. When you’re moving fast, copy-paste becomes muscle memory. Now I try to make a quick review part of the workflow, not an optional extra.

2. Give AI only what it needs.

Yes, context can improve results. No, that doesn’t mean you should hand over everything. When you can, strip names, email addresses, account details, and anything that identifies a person, client, or internal system. 

You can still use AI for real work. Instead of uploading a full confidential document, pull out the few relevant details and rewrite them in a non-identifying way so the tool can help without seeing sensitive content.

3. Be extra cautious with client and customer information.

Having access to information is not the same thing as having permission to upload it somewhere else. Handle client, customer, employee, and partner data with added care, and keep confidentiality terms or contractual restrictions front of mind.

4. Know the AI tool you’re using.

Don’t assume every AI product treats data the same way. Policies differ across tools, and can even differ between account tiers inside the same tool.

ChatGPT is a good example. On personal Free, Plus, and Pro accounts, chats might be used to improve OpenAI’s models unless you disable that in Data Controls. With ChatGPT Business and Enterprise, workspace data isn’t used to train OpenAI’s models by default.

So it’s not only “which tool am I using,” it’s also “which version of that tool am I using?

5. Check your privacy settings.

Many AI tools offer settings that affect how your information is saved, used, and shared. Take the time to look, especially before using an AI platform on work tasks. Defaults aren’t automatically the right choice for your organization.

Also, watch the Share button. In a lot of tools, sharing a chat produces a link that lets someone else view a snapshot of the conversation. Before you send that link, look back at what’s included. A thread that ends with something harmless can still contain client details, internal context, or sensitive notes earlier on.

The broader point is simple: know what will happen to your information before you put it in, and before you pass it along.

‍6. Create a company AI policy.

You probably don’t need a 40-page manual. But people do need clear ground rules. Decide which AI tools are approved, and spell out what can, and can’t, be entered.

For instance, brainstorming, proofreading, outlining, and high-level analysis might be fine. Client documents, contracts, internal financials, customer data, or anything unreleased might need approval. Passwords, credentials, highly sensitive personal details, and anything you’re contractually barred from sharing should never be included.

7. Keep a human in the loop.

Privacy isn’t the only reason to stay alert. AI can be wrong, miss key context, or present shaky information with total confidence. Read and verify what it produces before you send it, publish it, or lean on it for a decision.

Use AI, but do it on purpose.

I use AI daily and don’t plan to stop. It helps me move faster, sort through ideas, and offload some of the redundant parts of the job. But the better these tools get, the more important it is not to let speed outrun judgment.

Know the tool. Notice what you’re handing over. Remove what it doesn’t need. Follow your company’s rules. And when you’re unsure, ask before you upload.

The strongest AI users won’t always be the ones who use it nonstop. They’ll be the ones who understand what to share, and what to keep back.

Until next time,
Ashley

‍