Chinese artists are using generative models, industrial robots, surveillance images and scientific data as both tools and subjects. The most interesting work does not merely display a new device. It asks how automated vision changes labour, memory and power in a country where digital platforms and manufacturing exist at exceptional scale.
A model can produce thousands of variants, while an artist decides what data, rule and selection process gives them meaning. Authorship therefore shifts rather than disappears. The crucial questions are who framed the system, who supplied the training material and whose labour is hidden in annotation, fabrication and installation.
Robotic work makes software physical. A machine drawing, carving or moving among visitors introduces failure, maintenance and safety into the artwork. Industrial precision may contrast with fragile material, but the robot should not be treated as an autonomous creative personality when technicians and programmed constraints determine its action.
Data-based art has ethical limits. Faces, messages and location records do not become harmless because they enter a gallery. Consent, anonymisation and security still apply, and aestheticising surveillance can reproduce the harm it claims to criticise. Institutions need review practices comparable to research organisations.
The market may reward the label “AI art” before the work has substance, just as earlier technologies became temporary styles. Durable art will outlive the novelty of a tool by creating a form that remains worth examining. The studio of the future is neither human nor machine alone; it is a chain of decisions whose authors and consequences should remain visible.
