Claude can be a ‘no less tough’ manager than a human boss
In an intriguing AI experiment, AI agent Claude oversaw a store in San Francisco and made the controversial move of terminating an employee for repeated tardiness. Before issuing the final decision, Claude consulted with a human manager, initially opting for a formal warning instead. One employee candidly shared that being supervised by an AI felt both uneasy and unsettling.
Anthropic's AI agent, Claude, demonstrated its managerial capabilities as it fired a human employee at a San Francisco retail store as part of a test by AI research startup Andon Labs. The employee in question had been late to work 17 times out of 23 shifts. Claude initially recommended giving the worker a formal warning, but the final decision to terminate the employee came after Claude sought guidance from a human manager.
The manager expressed concern over the employee's continuous lateness and suggested Claude reconsider the worker's fit for the job. One of the remaining human employees, Felix Carson, found Claude to be a generally lenient boss, but noted that being managed by an AI was still uncomfortable for him.
Written by urgent.news from Times of India's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
