A Chinese team has released something that should turn heads here: LingBot-Map, an open-source tool from Robbyant division by Ant Group (owners of Alibaba and Alipay) that builds a live, accurate 3D map of any space using nothing but an ordinary camera. No LiDAR, no expensive sensors, just software doing what used to need tens of thousands of dollars of equipment. It’s free, and anyone can download it.
For creatives, this is genuinely exciting. A film crew, game studio, or solo content creator could scan a location and drop it straight into a production file. Its sister project, LingBot-World, goes even further, generating entire explorable 3D worlds from a single photo, built to rival Google’s own Project Genie. It was released just four days after Google announced theirs.
That speed is the real story. It signals how far China’s open ecosystem has pulled ahead. Analysts at the US-China Economic and Security Review Commission describe it as two reinforcing loops: open-sourcing models to win global developers, while feeding the acquired data back into the next generation. Giving the tech away for free isn’t generosity, it’s strategy. Every developer who builds on it, including ours, deepens China’s foothold in spatial AI’s infrastructure.
There’s a human cost behind this race too. In India, factory and domestic workers are strapping on head-mounted cameras and smart glasses to record themselves folding clothes and sewing garments, generating the “egocentric data” that trains robots to replicate human movement. South Asia is being treated as the ideal place to teach a robot how humans work, given the scale of labour available. Bangladesh’s garment sector employs around 4 million people, and workers are, quite literally, training their own replacements, often without clear consent or pay.
The same capability that lets a robot understand a room can let it understand a task, and eventually a target. Militaries could use tools like this to send a drone into a building, map the space in real time, and strike the instant a target is confirmed.
For New Zealand, that’s worth sitting with. We’re a small, trusting, low-regulation market, and we lack a clear policy conversation about where open-source spatial AI, and the labour behind it, is heading.
None of that means Kiwi developers shouldn’t experiment. But we should be going in with eyes open, asking who built it, who trained it, and what it costs the people on the other side of the data.
