Follow the latest coverage, related explainers and connected technology stories.
A test conducted by White Circle showed that an automated attack method was able to bypass the filters of DALL·E 3, Nano Banana 2, and GPT-Image-2 to produce explicit images, while newer releases from OpenAI and Google appeared more vulnerable according to the researchers. The findings reignite the debate over the need for mandatory, independent safety testing for AI models.
Circuit Breaker Labs tests artificial intelligence models through agents that simulate users of different ages, cultures, and languages, with the aim of uncovering interactions that could lead to psychological harm. The company is currently targeting AI-powered coaching, journaling, and mental-health support applications.
Anthropic is collaborating with Accenture to conduct independent evaluations and penetration testing of advanced AI models from within the company, with each party expected to invest at least $1 billion over five years. The initiative reveals a new model of oversight, but it still lacks stable standards for access, funding, and reporting.