Oracle宣布禁止开发者向OpenJDK开源项目提交AI生成的代码，原因涉及安全、保安和知识产权风险[1]。根据这一新政策，开发者可以在私下场景中使用大语言模型进行调试和代码审查，但不得向项目库、拉取请求或其他渠道上传AI生成的内容[1]。
这项禁令与Oracle的对外宣传形成了对比。公司高管此前曾表示AI模型已在为Oracle编写代码，其中首席执行官Larry Ellison声称该公司的AI模型正在执行此类工作[1]。公司副总裁Mike Sicilia也曾表示，AI工具使更小规模的工程团队能够更快地交付成果[1]。
Oracle has prohibited the submission of AI-generated code to the OpenJDK open-source project, citing security, intellectual property, and confidentiality risks [1]. While developers are permitted to privately use large language models for debugging and code review purposes, they are barred from submitting AI-generated content to project repositories, pull requests, or other submission channels [1].
This policy stands in contrast to Oracle's internal operations, where executives have previously claimed that the company's AI models are already writing code for Oracle [1]. Larry Ellison stated that Oracle's AI model is authoring company code, while Mike Sicilia asserted that AI tools enable smaller engineering teams to deliver faster [1]. Meanwhile, S&P downgraded Oracle's credit rating to BBB- (one notch above junk status), citing uncertainty about investment returns, even as the company invested 70 billion dollars in data center expansion this year [1].