The "Petri" tool deploys AI agents to evaluate frontier models. AI's ability to discern harm is still highly imperfect. Early tests showed Claude Sonnet 4.5 and GPT-5 to be safest. Anthropic has ...

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now Gabriel Chua, a data scientist at Singapore ...

TransferEngine enables GPU-to-GPU communication across AWS and Nvidia hardware, allowing trillion-parameter models to run on older systems. Perplexity AI has released an open-source software tool that ...

The tool, created by university researchers, is designed to find and automatically create a patch for vulnerabilities in large repositories like GitHub, but it isn’t perfect yet. Dutch and Iranian ...

Cloud-based cyber-attacks saw a marked increase in 2024, with threat actors adopting new tactics to exploit cloud resources at an unprecedented scale, according to Sysdig Threat Research Team’s (TRT) ...

Security researchers at Calif.-based Proofpoint have uncovered a large-scale account takeover campaign aimed at Microsoft Entra ID environments. The attackers are using TeamFiltration, an open source ...