Adanonymizer is a novel LLM plug-in that empowers users to navigate the trade-off between privacy protection and output quality during conversations with AI systems.

• Survey revealed a privacy paradox: users often disclose sensitive information despite knowing the risks
• The tool provides an interactive trade-off curve allowing users to control their privacy-performance balance
• Enables context-aware anonymization of personal information in real-time
• Addresses critical security concerns by giving users agency over what personal information LLMs can access

This innovation is particularly significant for security professionals as it offers a practical solution to mitigate privacy risks in enterprise LLM deployments without sacrificing the utility of AI assistants.

Adanonymizer: Interactively Navigating and Balancing the Duality of Privacy and Output Performance in Human-LLM Interaction