ChatGPT and Claude, among the latest large language models, performed worse than expected in the Stroop test, a psychology experiment used to gauge human attention and executive control, a study found. Researchers tested OpenAI\'s GPT-4o and Anthropic\'s Claude 3.5 Sonnet and saw performance fall sharply when word meaning conflicted with ink colour, with accuracy declining further as tasks increased. Follow-up tests on newer models showed limited improvement, the study said.

Cisco unveiled its integrated operations platform, Cisco Cloud Control, and a suite of security products as AI agents spread inside companies. Announced at the Cisco Live event in Las Vegas, the platform combines network, security, observability, infrastructure and collaboration management in a single operating environment. Cisco also introduced Cloud Control Studio and Cisco AI Canvas. In security, it unveiled Live Protect and strengthened AI agent governance with DefenseClaw and other features, adding Splunk enhancements and pledging quantum-safe support across much of its portfolio by end-2026.