Moxie Marlinspike, creator of Signal Messenger, is launching Confer, an open-source AI assistant designed to ensure user data remains private and unreadable by anyone except the account holders. Utilizing strong encryption and trusted execution environments, Confer aims to set a new standard for AI chatbots while maintaining user confidentiality and security.

Lattica has unveiled a new platform utilizing fully homomorphic encryption (FHE) to allow AI models to process encrypted data without exposure. The company secured $3.25 million in pre-seed funding to enhance the security and privacy of AI applications. This innovative approach enables AI providers to host and manage models while ensuring that sensitive data remains protected.

Using HashiCorp Vault’s transit secrets engine, sensitive information in Amazon Bedrock Knowledge Bases can be encrypted, particularly when utilizing retrieval augmented generation (RAG). The demo illustrates how to process vacation rental data while ensuring personal identifiable information (PII) remains secure and undisclosed.

1Password emphasizes the importance of security in AI integration, outlining key principles to ensure that AI tools are trustworthy and do not compromise user privacy. The principles include maintaining encryption, deterministic authorization, and auditability while ensuring that security is user-friendly and effective. The company is committed to creating secure AI experiences that prioritize privacy and transparency.