The risks

When software takes on a body, the risks become tangible: accidents, jobs, surveillance, war, geopolitical dependence.

Industrial history solved safety with the cage: separating robots and humans. Cobots, humanoids and service robots remove that separation. Certifying a learned (and therefore non-deterministic) robot among humans remains a largely open problem.

The first wave of industrial robots had real local effects on jobs and wages. The humanoid wave would mostly affect manual service jobs. On the other side: massive demographic shortages. The net balance will depend on transition policies.

A hacked robot is not just a compromised computer: it is a mobile object, fitted with cameras, sometimes heavy. A poorly secured fleet can become a physical botnet. At home, teleoperation and cameras raise a direct question: who can see inside?

Drones, loitering munitions, autonomous systems: Ukraine has accelerated the robotisation of war. The debate on lethal autonomous weapon systems is stalling at the UN while the technology advances.

Rare earths, magnets, batteries, lidars, low-cost components: the value chain leans towards China, which can use its export controls as an industrial weapon.

Anthropomorphism is powerful: attachment of children or elderly people, care robots replacing human presence, emotional manipulation. The law is starting to follow, but usage will move fast.