Can I sell my AI chat data? Why would someone buy it?

If you’ve ever asked ChatGPT something a little too personal — a health worry, a relationship problem, a draft email written in the heat of the moment — the idea of sharing that conversation can feel uncomfortable. Very understandable.

When we speak to people, most people’s initial reaction is: who would even want that, and what do you do with it?

This post walks through exactly what happens when you share data through WeBuyData, and answers those two questions.

What “selling your data” actually means here

When people hear “we buy your data,” they often picture something invasive — like someone reading your private messages. That’s not how this works.

In practice, contributing through WeBuyData means:

• You choose what to share (specific chat exports, usage patterns, or other digital activity — never your full device or account access). There is a whole category of vendors out there called “clickstream data” providers who do exactly that. These are firms such as MFour, Datos and BigDBM. We do not operate in this category — and it typically offers very low incentives to contributors.
• Identifying details are removed or separated from the content before it’s used.
• You’re compensated for the contribution, not for ongoing surveillance.
• You can see what you’re agreeing to before you submit anything.

Lastly, researchers need many hundreds of datasets and tens of thousands of messages to draw meaningful conclusions. Even if it were useful, it would not be feasible for any researcher to review individual contributions in detail. They are looking for aggregate trends in real data to support their thesis.

TL;DR — your messages are only valuable in the aggregate, and nobody is going to go through them one by one.

Over 7,000 contributors have shared data through WeBuyData since launch.

Why does anyone want this data in the first place?

AI companies, researchers, and product teams need real examples of how people actually use AI tools. While researchers often use synthetic conversations generated by AEO (answer engine optimisation) companies, actual conversations are much more useful. Right now, most of that information either isn’t collected at all, or it’s collected by the AI platforms themselves without any direct compensation to the person who generated it.

WeBuyData exists because we think that’s backwards. If your data is valuable enough for a company to want it, you should be the one getting paid for it.

What we do and don’t do with your data

To be transparent:

We do:

• Anonymise or pseudonymise contributions before they’re used
• Tell you upfront what a specific data request involves
• Pay you for what you choose to share
• Let you decline any individual request

We don’t:

• Sell your name, email, or contact details alongside the content
• Access your accounts directly or request your passwords
• Require you to share anything you’re not comfortable with

Is this even legal?

Yes. Selling or licensing your own data, with informed consent, is a legitimate and growing part of the data economy — it’s the same underlying principle that lets you opt in to loyalty programs, browser extensions, or research panels that compensate you for sharing usage data.

For example, millions of people use the Chrome extension Honey to get free discount codes when shopping. The trade-off is that Honey gets access to your browsing data (which people often don’t realise) — and it’s why Honey was sold to PayPal for $4 billion in 2020.

The difference with WeBuyData is transparency: you know what’s being asked for, and you’re paid directly for it.

The honest trade-off

We don’t pretend there’s no trade-off at all. Sharing any data involves some level of exposure, even when it’s anonymised. The question worth asking yourself is the same one you’d ask about any data-sharing decision: am I comfortable with what’s being asked for, do I understand how it’ll be used, and am I being fairly compensated for it?

If the answer is yes, we’d love to have you. If it’s no, that’s completely understandable.