Google Secure AI Framework (SAIF)

Why it matters

• It gives security teams a shared vocabulary for AI-specific threats that conventional application security programmes do not cover.
• It works alongside the OWASP Top 10 for LLM Applications (specific weaknesses) and ISO/IEC 42001 (the management system) as the security layer of an AI programme.
• It is voluntary: no regulator requires it, but it documents a secure-by-design approach in customer and vendor reviews.

In detail

A voluntary industry framework, not a law or a standard. Its six core elements are: expand strong security foundations to the AI ecosystem; extend detection and response to bring AI into the organisation's threat universe; automate defences to keep pace with existing and new threats; harmonise platform-level controls for consistent security across the organisation; adapt controls to create faster feedback loops for AI deployment; and contextualise AI system risks in surrounding business processes. Google notes the elements are not meant to be applied in sequence.

Related concepts

• OWASP Top 10 for LLM ApplicationsThe OWASP Top 10 for LLM Applications is the Open Worldwide Application Security Project's list of the ten mos...
• Prompt injectionPrompt injection is an attack in which crafted input, typed by a user or hidden in content the model reads, ov...
• Data poisoningData poisoning is an attack that corrupts an AI model by inserting or altering its training or fine-tuning dat...
• AI systemA machine-based system that operates with some autonomy, may adapt after it is deployed, and infers from the i...
• Red teaming (AI)AI red teaming is structured adversarial testing in which testers try to make an AI system fail, by producing...