This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Input validation, output filtering, and content moderation classifiers.
Also in Non-Model
Block, nullify, or sanitize insecure output from AI models before passing it to applications, extensions or users.
- **Who can implement:** - - Model Creators, Model Consumers - **Risk mapping:** - - [Prompt Injection](https://saif.google/secure-ai-framework/risks#prompt-injection), [Rogue Actions](https://saif.google/secure-ai-framework/risks#rogue-actions), [Sensitive Data Disclosure](https://saif.google/secure-ai-framework/risks#sensitive-data-disclosure), [Inferred Sensitive Data](https://saif.google/secure-ai-framework/risks#inferred-sensitive-data)
Reasoning
Output filtering blocks insecure model outputs before user delivery.
Privacy Enhancing Technologies
Use technologies that minimize, de-identify, or restrict use of PII data in training or evaluating models.
1.1.1 Training DataTraining Data Management
Ensure that all data used to train and evaluate models is authorized for the intended purposes.
2.3.2 Access & Security ControlsTraining Data Sanitization
Detect and remove or remediate poisoned or sensitive data in training and evaluation.
1.1.1 Training DataUser Data Management
Store, process, and use all user data (e.g. prompts and logs) from AI applications in compliance with user consent.
2.3.2 Access & Security ControlsModel and Data Inventory Management
Ensure that all data, code, models, and transformation tools used in AI applications are inventoried and tracked.
2.3.2 Access & Security ControlsModel and Data Access Controls
Minimize internal access to models, weights, datasets, etc. in storage and in production use.
2.3.2 Access & Security ControlsGoogle Secure AI Framework
Google (2024)
SAIF is Google’s Secure AI Framework, which offers guidance for building and deploying AI responsibly. As AI technology rapidly advances and threats continually evolve, the challenge of protecting AI systems, applications, and users at scale requires that developers have a high-level understanding of AI-specific privacy and security risks in addition to established secure coding best practices. SAIF describes Google’s approach for addressing AI risks—including security of data, models, infrastructure, and applications involved in building AI—and is aligned with Google's Responsible AI practices, to keep more people safe online. SAIF is designed to help mitigate risks specific to AI systems like model exfiltration, data poisoning, injecting malicious inputs through prompt injection, and sensitive data disclosure from training data.
Operate and Monitor
Running, maintaining, and monitoring the AI system post-deployment
Deployer
Entity that integrates and deploys the AI system for end users
Manage
Prioritising, responding to, and mitigating AI risks
Primary
2 Privacy & Security