بچه‌های آنتروپیک یه سیستم جدید و کاردرست به اسم FlashProve معرفی کردن که با متدهای رمزنگاری، مو لای درزش نمیره و ثابت می‌کنه جوابی که گرفتی واقعاً از شست پای همون مدل هوش مصنوعی دراومده یا نه

New Anthropic research: progress on provable inference for large language models. Here, we introduce FlashProve. Building on established cryptographic techniques, FlashProve allows developers to verify whether a given response came from a given AI model. It’s substantially more efficient than prior open-source methods for doing so. These proofs have a range of applications. They can, for instance, make it harder for attackers to hijack a chat, or enable safety evaluators to verify that the model they’re testing is the same one being served to customers. For our full write-up, see here: https://www.anthropic.com/research/provable-inference Further progress in provable inference will likely be an important part of improving trust in AI deployment. To that end, we're open-sourcing FlashProve’s code, here: https://github.com/anthropics/FlashProve We’re grateful to our collaborators: Alessandro Chiesa, @GiacomoFenzi, @gwberg95, @miranda__christ, Rahul Dalal (http://rdalal.github.io) and Ben Mayer (@transmissions11).