MIRI这篇论文列了28种监控AI研究的方法，连搜家、测谎、监禁都写出来了，挺硬核的，想了解AI监管走向的建议看看。
机器智能研究所（MIRI）发布论文，主张对AI研究本身进行限制。论文列出了28种管控机制，包括代码审查、员工监控、芯片使用监控和国际搜查令。机制还涵盖测谎、模型允许列表以及可能涉及监禁的惩罚措施。该论文旨在建立全面的研究控制制度来监控研究者和机构。
Sure, why not.
Sure, why not. Nirit Weiss-Blatt, PhD @DrTechlash The Machine Intelligence Research Institute calls for restricting AI research itself. It lays out a research-control regime for monitoring researchers and organizations, including penalties that "could plausibly include prison sentences." This new paper catalogs 28 mechanisms, including intelligence gathering, international search warrants and inspections (of properties, computers, and files), polygraphs; inference-content monitoring (of user prompts, tool use, model outputs), sting operations, AI-assisted code review (which may extend to codebases, experiment logs, and research documents), embedded auditors, employee-monitoring software, chip-use monitoring, training-data review, and model allowlists. 🔗 View Quoted Tweet 💬 17 🔄 1 ❤️ 33 👀 17902 📊 12 ⚡