Waiting for Anthropic Support (koenvangilst.nl)

A user recently experienced a significant issue with their Anthropic Claude API key, which was compromised due to a breach at OpenCode, leading to unexpected charges on their account. Upon attempting to seek support, the user encountered a frustrating AI-driven customer service system that failed to adequately address their issue regarding the compromised key. Although the chatbot directed them to human support, it took several days to receive a response, highlighting inefficiencies in the AI's ability to categorize support issues effectively. This situation underscores the challenges and limitations of AI in managing customer support, especially in the context of sensitive incidents like security breaches. While Anthropic has integrated AI with traditional customer service, the experience reveals a gap in the technology's effectiveness in resolving complex issues in a timely manner. This incident raises important questions for the AI/ML community regarding the scalability and reliability of using AI for customer support and how companies can better leverage AI to enhance personal interactions while maintaining security standards.