AI Risk Library
How risky is your team's AI?
A reference library of 500+ AI tools, reviewed across 9 attributes that matter to security and procurement teams.
500+ AI tools · 9 risk attributes · Free, no signup required
Explore up to 5 AI tools, then book a meeting for the full industry analysis.
How the Risk Library works
9 attributes that matter to security and procurement teams.
Each tool in the library is reviewed against nine attributes using publicly available evidence: vendor documentation, terms of service, privacy policies, and regulatory disclosures. The Risk Library is a high-level screening aid, not a security audit.
01
Audit log / trail
Whether the tool produces an audit log of user activity, what's captured, and whether logs are exportable.
02
Geographic data centers
Where the tool processes and stores data. Critical for data residency requirements in regulated industries.
03
Intellectual Property
Whether the tool's terms of service preserve customer IP ownership of inputs and outputs, or claim license over them.
04
Prompt injection
Whether the tool has known vulnerabilities to prompt injection attacks, and what mitigations are in place.
05
SOC 2 / ISO 27001
Compliance certifications the tool holds. Required for many enterprise procurement processes.
06
Supply chain vulnerabilities
Disclosed CVEs, security incidents, and breach history affecting the tool or its dependencies.
07
Third-party AI risks
Which third-party AI models or services the tool depends on, and what data flows to those parties.
08
Training on data
Whether the tool trains its models on customer prompts and outputs by default, and whether opt-out is available.
09
Zero data retention
Whether the tool offers a zero-data-retention mode where inputs and outputs are not stored after processing.
The bigger picture
This library tells you about a tool.
Risk Monitoring tells you what your team is actually using.
Looking up one tool is a start. Knowing what your employees use, who's using it, and with what data, is the real question.
01
What this library does
Helps you assess a specific AI tool before deciding whether to allow it. One tool at a time, in any browser.
02
What this library can't tell you
Which AI tools your employees are using right now. Whether they're sharing sensitive data. Which department is using what. What's been growing in usage over the last 30 days.
03
What Risk Monitoring does
Continuous, on-device visibility into every AI tool, account, and prompt across your organization. Same 9-attribute risk model, applied to your real environment.
Trusted by
Trusted by security teams evaluating real AI risk.
Customers & Partners
Stop assessing one tool at a time.
Get started
See every AI tool, account, and prompt across your organization in real time. Continuous visibility into your real environment.
Questions? sales@magicmirrorsecurity.com
