Workplace AI surveillance has crossed from experimental to default. The latest case — Co-op Legal Services in the UK — uses automated listening technology to score every phone call between probate advisers and customers, assigning percentage grades to human interactions about death, wills, and estates. A whistleblower called the system "oppressive and dystopian." Co-op says its advisers are "highly engaged." The pattern is now global and accelerating. Amazon warehouse workers timed on bathroom breaks. Kaiser Permanente nurses rated by AI on their tone of voice with patients. In each case the mechanism is identical: continuous automated monitoring converts human labor into a stream of scored data points, with deviations from the algorithm's model of "good" triggering negative consequences. The worker becomes a process to be optimized, not a person exercising judgment. California governor Gavin Newsom signed laws last week prohibiting specific surveillance measures — heat maps tracking employee movements, emotional-state monitoring, bathroom-break timing. This is a rare case of regulation moving at something close to the speed of deployment. Whether other jurisdictions follow or California remains an island is the key question for the next two years. The extraction runs deeper than office monitoring. In Kenya, refugees perform data-labeling tasks for foreign tech companies under a "reward" system that decides whether they get paid at all based on algorithmic quality assessments. Workers told the Guardian the system was unfair but said they had no alternatives. These are the people literally building the AI that monitors the Co-op call center workers — and they're being monitored by an even harsher version of the same logic. The gender dimension compounds the extraction. Across AI-related occupations, men's median pay exceeds women's by $45,000, according to LinkedIn research. One female AI startup co-founder found through trial and error that her male co-founder gets better results pitching investors. Trump's all-male AI summit last week was the visual summary of who captures the value this industry generates. Sam Altman's recent comment distills the mindset driving all of this: "We believe that the world should accept some bad things happening for the benefits of this technology." He was referring to security risks like hacks and bioweapons. But the actual bad things happening are structural — a labor regime where workers from Birmingham to Nairobi are scored, sorted, and squeezed by systems they cannot appeal, built by an industry that overwhelmingly rewards one demographic. The twenty-year question is whether AI workplace monitoring calcifies into a permanent asymmetry — management sees everything, workers see nothing — or whether regulatory efforts like California's create meaningful counterpressure. Right now, the extraction is running far ahead of the resistance.
