Has anyone actually used ChatGPT with real financial data?

UK General Chat Tech & Gaming
,
1

Hi all,

I’ve been wondering about something lately. With tools like ChatGPT becoming more common, I’ve noticed people using them to draft complaints, understand transactions, or even summarise financial emails.

What I’m curious about is what people actually do in practice when it comes to their data. Do you take the time to remove sensitive details before pasting things in, or do you just use it as-is?

I get the feeling that most people probably don’t think too much about it in the moment, even though it involves pretty personal information.

Would be really interesting to hear how others approach this in real life.

2

I would never do it in the first place. Using AI would make me feel like I’d failed as a human being!

5 Likes
3

Seriously, though, I find it quite horrifying how easily people are giving over their cognitive abilities to machines.

2 Likes
4

Exactly. I think it’s less about replacing thinking and more about how people use it. What concerns me more is how casually people share sensitive info with these tools.

I’ve seen people paste full financial details without thinking twice.

Feels like the bigger issue is how data is handled, not the tool itself.

5

I have no issue with AI, but I do have limits on what I’ll do.

Generally I give it snippets of code that I’m working on where my syntax is off (I’m learning) and it will correct it, things like that are invaluable.

Yesterday I was making a slide deck and couldn’t think of the wording for a title for what I wanted to convey. I asked and I was given great suggestions.

But I’m not handing over my financial data. I don’t really care about the privacy/security aspect of it, there’s just not anything AI can do with it that I can’t myself.

1 Like
6

That makes sense, and I think most people use it that way.

The interesting part is where that line is — code or wording feels fine, but financial data and other sensitive data is different.

What I’ve noticed though is not everyone draws that line so clearly.

7

Oh no they definitely don’t.

We have company Copilot licenses and it’s shiiiiiiiiite! So I try to stick with ChatGPT but then I have to be a lot more sensitive about what I ask/share.

@coffeemadman had AI ask him about his day out with his partner, without the partner being mentioned, AI had just presumed they were together based on previous conversations! It’s very aware and this is only the beginning.

2 Likes
8

Erm, you know Copilot is built off Chat GPT right? So if you toggle the work/web slider in Copilot it (should) give a comparable experience?

And to answer the OP question absolutely not. While I trust the guardrails are in place for most part with consumer AI (so, not OpenClaw :joy:) I didn’t and still don’t like giving any database more info than it needs so it can then shuck products to me based on that.

1 Like
9

Yeah, that makes sense.

It’s useful, but you end up being much more careful about what you share. That example you mentioned is exactly what makes it feel a bit unpredictable.

Feels like the awareness is there, but the boundaries aren’t always clear yet.

10

Yeah, that’s a good point.

Even with guardrails, it still comes down to how much you trust where the data is going and how it’s used. I think that uncertainty is what makes people cautious more than anything else.

1 Like