What is egocentric (first-person) video — and why AI companies pay for it

"Egocentric" is just a fancy word for first-person video: the camera sees roughly what your eyes and hands see. It happens to be some of the most valuable data in AI right now, and recording it is one way everyday people earn money helping train these systems.

Egocentric vs. ordinary video

Almost all the video online is shot from the outside, pointed at a scene. Egocentric video sits inside the action instead. Your hands are in the frame, handling real objects, the same way a robot or an AI assistant would have to see the world before it could act in it.

That first-person, hands-visible angle is the piece most existing datasets are missing. Scarcity is what makes it worth paying for.

Why AI and robotics labs need it

A robot has to learn how a physical task actually unfolds: the order of the steps, how an object gets gripped, what goes wrong, and how a person recovers from it. Thousands of real examples filmed from the doer's point of view teach that better than anything else.

You can't pull this off the web at the quality required, so labs pay real people to film it. Done properly, that comes with your consent and clear terms for how the footage gets used, and you can withdraw that consent later.

What recording it looks like for you

You film ordinary activities like cleaning, cooking prep, organizing, or yardwork, with your hands visible, in landscape, at a decent resolution. Each task type comes with a short guide telling you how to frame it.

It's the same stuff you'd be doing around the house anyway. The only real change is that a camera is rolling so an AI can learn from it.