Figure AI’s Index App Claims 16 Million Robot Training Videos and 264,000 Downloads

Imagine being paid to film yourself cooking dinner, stacking tins in a cupboard, or fixing a leaky tap. That’s essentially what Figure AI is asking people around the world to do — and according to the company, hundreds of thousands of them have already signed up.

Figure AI, a US-based humanoid robotics start-up, has publicly launched Index, a mobile app that pays contributors to record themselves performing everyday physical tasks. Those videos are then used to train the company’s robots to understand and replicate real-world human activity. After operating quietly for several months, Figure AI announced the platform’s scale in a post on X in late August 2026 — and the numbers are striking.

What Is Index and How Does It Work?

The idea behind Index is straightforward, even if the technology behind it is anything but. Traditional robot training has leant heavily on computer simulations or tightly controlled lab environments. Figure AI’s argument is that a robot capable of working in a real home or warehouse needs to learn from real human behaviour — not a virtual version of it.

So the company built Index. Approved contributors, called “Creators”, use the app to record short videos of themselves doing ordinary things: cooking, cleaning, using tools, stocking shelves, making repairs. Figure AI says it is currently ingesting around 30 minutes of uploaded video every single second. That’s a vast amount of data arriving continuously, around the clock.

According to Figure AI’s own announcement, Index has now been downloaded around 264,000 times across roughly 108 countries, with about 44,000 weekly active users actively contributing footage. The total video count stands at around 16 million uploads. The company describes Index as “the largest and most diverse robot dataset to scale general purpose robots” — though it’s worth being clear that this claim has not been independently verified against every other robot dataset in existence. Several AI researchers and policy commentators have called for more transparency on how the dataset is composed and evaluated.

Feeding the Figure 03

All of this data has a specific destination. Figure AI says Index feeds directly into Helix, the AI control system that drives its Figure 03 humanoid robot. The goal is a general-purpose machine capable of switching between tasks — not a robot that can only do one thing in one setting.

The company has published some internal metrics about what’s inside the dataset. By their own figures, every 1,000 hours of Index data contains around 373 different kinds of tasks, roughly 1,146 distinct objects, and footage across 116 different environments. These are company-provided numbers and haven’t been validated by outside researchers, so they should be read as illustrative rather than definitive.

Figure AI has also reported paying contributors around $15 million to date, with plans to spend around $1 billion over the next 12 months on data collection and computing. Both figures are in US dollars — around £12 million and £790 million respectively at current rates — and come from the company’s own communications rather than any independently audited source.

Privacy, Pay, and the Bigger Picture

Not everyone is enthusiastic. Critics have raised concerns about what it means to hand over large volumes of video footage — filmed in your own home, your workplace, your daily life — to a private American company building commercial robots.

UK data protection law, including the UK GDPR and the Data Protection Act 2018, would apply to any UK residents using Index, requiring the company to be transparent about how personal data is used, stored, and shared. The Information Commissioner’s Office has not issued any specific guidance on Index at the time of writing, and no UK government body has commented on the platform directly.

Labour advocates raise a separate concern. Training robots to perform physical tasks at this scale could, over time, accelerate automation in sectors like logistics, retail, warehousing, and care work. The people contributing videos today could, in theory, be helping to train the machines that eventually take on roles similar to their own.

Brett Adcock, Figure AI’s chief executive, has framed Index differently, describing it as a way to democratise access to robot training and build a dataset broad enough to make robots genuinely useful in the real world. The company positions the Creator model as a legitimate income opportunity for participants.

But some gig-economy researchers point out that variable acceptance rates, task selection, and earnings uncertainty make such platforms an unreliable source of income for workers who come to depend on them.

What This Means for Kent Residents

There’s no evidence that any Kent employer, public body, or institution is currently involved with Index. But for anyone in the county with a compatible smartphone, the app is — in principle — accessible, and contributing videos is a potential source of supplementary income if you’re accepted as a Creator. More broadly, Kent workers in warehousing, retail, logistics, and facilities management may want to keep an eye on how platforms like Index accelerate the development of physical robots, since those are exactly the sectors where general-purpose humanoid machines are most likely to appear first.