Anthropic's AI 'Claudius' Experiment as Retail Store Manager Results in ₹60 Lakh Loss

A unique experiment conducted by research company Anthropic in the US put an AI model named 'Claudius' (based on Claude) in charge of a retail store as its manager, giving employees the option to rest and consult the AI for any issues.

Key Highlights:
- The Experiment: The AI managed inventory, pricing, orders, and staff instructions for roughly four months to test autonomous store management.
- Operational Flaws: The AI made poor logistical decisions, over-ordered some items while missing essential stock, forgot its social media account passwords, and struggled with business profitability versus customer appeasement.
- Financial Impact: The four-month test resulted in a loss of approximately ₹60 lakhs, proving that AI is not yet ready to completely replace human managers and is best utilized as a supportive assistant.