AI models that have a pro-China lean may systematically generate more vulnerable code when writing applications for Western organizations, cybersecurity firm Hirundo has discovered. The […]