Elon Musk and co may relish march of the robots but there must be AI boundaries in the workplace | Heather Stewart
TL;DR
Heather Stewart opens with D4YRL, a robot magician rejected by the Magic Circle: the tricks worked, but the machine failed at the human part of performance. Sarah O’Connor’s book ‘We Are Not Machines’ supplies the reporting: surveilled Amazon work, hidden AI trainers, and translators pushed into cheaper post-editing of AI output. The argument is not anti-AI. It is about choosing where automation belongs, especially in care, creative work and systems that monitor employees.
Nauti's Take
AI should make work lighter. When it pushes people into cheaper correction loops or treats care, creativity and attention as replaceable process steps, the boundary has already been crossed.
Companies have little incentive to draw that line on their own because the short-term efficiency gain is too tempting. Worker negotiation and clear public rules need to sit at the start of major AI rollouts, not after the damage is visible.
Briefingshow
The core issue is power: who gets to decide which work is automated and which human qualities are treated as expendable? If workers only get a say after deployment, AI becomes a management tool first. It can cut costs, intensify work and flatten creative or caring roles into metrics that miss the real value.
Sources