HN Explorer

Ask HN: Why would a company ban ChatGPT?

3 points by zw123456 ↗ HN
I work for a large corporation, that I don't feel comfortable naming, but they sent out an email to all 160K employees forbidding anyone from using Bing, any ChatGPT, GPT of any kind on company computers for security reasons. They provided no explanation other than saying "for security reasons". Does anyone have any idea what it is they are worried about? I don’t see where it could be an issue. How would it be any less secure than a regular search engine?

9 comments

[ 4.5 ms ] story [ 37.4 ms ] thread
mdmglr
An abundance of caution?

If employees start sharing proprietary data. Could even be data you consider harmless like quarterly results before the earnings meeting. Could be a legal issue.

I think from an OpSec perspective band the use then come up with a policy.

adamquek
They are afraid that you will share personal or proprietary data with OpenAI. Not everyone is trained in handling with sensitive information, and it may be tempting for new tech users to just put in data and ask ChatGPT to do analytics for them, which can be a potential threat.
dschuessler
There are interesting technical solutions for preventing sensitive informations from leaking that could probably enable a somewhat safe use of ChatGPT and the likes in a corporate context.[0]

[0]: https://www.cloudflare.com/products/zero-trust/dlp/

verdverm
How is ChatGPT going to provide a response? It has to read the plain text.

Corporations are worried that OpenAI will use their proprietary data at some point, which might end up being memorized and regurgitated later.

dschuessler
I think the idea behind these products is that network requests and responses containing sensitive data never reach their target in the first place. So if someone tries to tell ChatGPT sensitive data, the outgoing request is simply blocked. Harmless content goes through, though.
verdverm
> Cloudflare Data Loss Prevention (DLP) secures sensitive data in transit.