China National Radio Exposes AI Telemarketing Systems Making 1,500 Harassment Calls Daily
China National Radio exposed the operations of AI telemarketing systems that can make 800 to 1,500 automated calls per day per robot. The report highlights a shift from traditional rule-based bots to advanced LLM-powered agents that dynamically generate scripts based on user emotions. This exposure reveals how low-cost AI technology is being weaponized for mass harassment, bypassing traditional telecom filters and operator tracking. It underscores the growing challenge of protecting consumer privacy and enforcing anti-harassment laws in the era of generative AI. The AI systems use disguised physical phone numbers routed through internal lines, leaving no call logs on the operator's system to prevent consumer complaints. Additionally, telemarketers utilize "one-click unmarking" services up to three times a day to clear spam labels applied by security apps.
## BACKGROUND
AI outbound calling systems automate telemarketing by integrating technologies like Automatic Speech Recognition (ASR), Text-to-Speech (TTS), and Natural Language Processing (NLP). While traditional systems rely on pre-recorded audio and rigid decision trees, newer LLM-based agents can engage in fluid, human-like conversations.