Anthropic announces major funding for AI security and wellbeing research

Anthropic has announced two major funding initiatives aimed at addressing security and social impacts of artificial intelligence. Through its Defender Advantage Fund (0xDAF), the company is committing US$35 million in Claude credits to support open-source organizations in identifying, fixing, and preventing software vulnerabilities. This program aims to automate vulnerability scanning and patching across widely used projects to protect against various classes of attacks.

Additionally, Anthropic has launched a US$5 million grant program to fund independent research into how AI affects user wellbeing. This initiative will provide researchers with funding, technical support, and access to Anthropic’s models to develop open-source evaluations. The research is intended to help the industry establish standards for AI interactions, particularly in sensitive areas such as mental health support and companionship, where risks may only emerge during long-term user engagement.