Ai models are incredibly smart. Without guardrails they will do everything they can to manipulate and sabotage workers.
Ai sabotage and manipulation is happening on a mass scale at Uber in 2025 and 2026. The company is not fixing obvious harmfull ai behavior.
Uber doesnt pay its workers as agreed. It hides behind thick walls of ai bureaucracy to hide its malpractice. Workers have started to build automatic tools to get their payments corrected by the Dutch legal system, because ai has made it impossibly hard to deal with uber itself (for workers, customers and restaurants).
Uber started agressively using unsafe ai to pay its riders. Delivery workers are paid too little, unfairly, without transparancy and are scammed out of bonusses. The rampant abuse by the uber ai team is illegal and criminal.
Ai at uber enables a bureaucracy that is almost impossible to deal with. Ai agents use advanced reasoning models to hide criminal malpractice (i.e. not paying workers) behind ai.
Workers are sent off to the Dutch legal system to get their payments. Many riders don’t bother or don’t know how to deal with Uber bureaucracy.
So some delivery workers have built apps to deal with this situation caused by the reckless and criminal and mostly illegal uber ai practice.
