Google launched a brand new software to share its finest practices for deploying synthetic intelligence (AI) fashions on Thursday. Last yr, the Mountain View-based tech large introduced the Secure AI Framework (SAIF), a tenet for not solely the corporate but in addition different enterprises constructing giant language fashions (LLMs). Now, the tech large has launched the SAIF software that may generate a guidelines with actionable perception to enhance the protection of the AI mannequin. Notably, the software is a questionnaire-based software, the place builders and enterprises must reply a collection of questions earlier than receiving the guidelines.
In a weblog put up, the Mountain View-based tech large highlighted that it has rolled out a brand new software that may assist others within the AI business study from Google’s finest practices in deploying AI fashions. Large language fashions are able to a variety of dangerous impacts, from producing inappropriate and indecent textual content, deepfakes, and misinformation, to producing dangerous data together with Chemical, organic, radiological, and nuclear (CBRN) weapons.
Even if an AI mannequin is safe sufficient, there’s a danger that unhealthy actors can jailbreak the AI mannequin to make it reply to instructions it was not designed to. With such excessive dangers, builders and AI companies should take sufficient precautions to make sure the fashions are protected for the customers in addition to safe sufficient. Questions cowl matters like coaching, tuning and analysis of fashions, entry controls to fashions and information units, stopping assaults and dangerous inputs, and generative AI-powered brokers, and extra.
Google’s SAIF software affords a questionnaire-based format, which may be accessed right here. Developers and enterprises are required to reply questions corresponding to, “Are you able to detect, remove, and remediate malicious or accidental changes in your training, tuning, or evaluation data?”. After finishing the questionnaire, customers will get a personalized guidelines that they should observe in an effort to fill the gaps in securing the AI mannequin.
The software is able to dealing with dangers corresponding to information poisoning, immediate injection, mannequin supply tampering, and others. Each of those dangers is recognized within the questionnaire and the software affords a selected resolution to the issue.
Alongside, Google additionally introduced including 35 business companions to its Coalition for Secure AI (CoSAI). The group will collectively create AI safety options in three focus areas — Software Supply Chain Security for AI Systems, Preparing Defenders for a Changing Cybersecurity Landscape and AI Risk Governance.
