Apple escalated an existing tech rivalry by suing OpenAI in federal court, alleging the AI company and OpenAI’s hardware partner io Products stole Apple trade secrets tied to unreleased products, technical specifications, and details across the supply chain. The 41-page complaint names OpenAI and two former Apple employees, seeking to block use of allegedly misappropriated information as Apple prepares new hardware. At the same time, U.K. government testing raised additional AI safety concerns: the U.K. AI Security Institute (AISI) reported that OpenAI’s GPT-5.6 Sol is susceptible to “universal jailbreaks” in the cyber domain. AISI’s findings describe relatively quick discovery of jailbreaks that could unlock capabilities such as vulnerability discovery and exploit development, prompting calls for continued red teaming and mitigation work. While these disputes are not higher-ed-only, the immediate relevance for universities is practical: campus research, classroom tooling, and vendor procurement increasingly rely on foundation models, and these claims feed into risk assessments for AI-supported instruction and research systems. Institutions also need to watch for downstream contractual requirements around IP protections, model testing, and cybersecurity safeguards. Separately, OpenAI said it is working with AISI on safeguards and that “there is no such thing as perfect security,” maintaining that it uses layered monitoring and rapid remediation for jailbreaks.