BLUE TRAIL AI SECURITY LAB
Created to address growing Al/ML risks and security concerns of Blue Trail Software's customers Combines cross functional teams from Blue Trail Software to apply the requisite skillset & technologies to ML risks and cybersecurity Maintains expertise, skills, best practice guidelines and ML security related research projects
AI MUST BE TRUSTED, SAFE AND SECURE
- Equitable algorithms (fairness, bias)
- Ethical Al (no 'off label' uses, etc.)
- Compliance, legal and reputational risks
- Have a full understanding of the Al decision-making processes
- ML monitoring
- ML auditability
- ML error analysis
- MLOps
- Equitable algorithms (fairness, bias)
- Ethical Al (no 'off label' uses, etc.)
- Compliance, legal and reputational risks
APPLYING INDUSTRY BEST PRACTICES TO
ML SECURITY ASSESSMENT
A Threat Model & Security Analysis (TMSA) is a key document used in the security industry to define and scope the security evaluation of a system, the target of evaluation:
- Description of the Assets in the system to be protected
- Use-cases in which the assets are involved
- Security measures in place to protect the assets
- Attacker's profile, Threat model and scenarios
The Security Assessment is based upon your selection of threat scenarios defined in the TSA, and leverages security and explainable Al tooling.
The Security Assessment Report lists tested Threat Scenarios (TSA), describes attack attempts, and list identified risks and vulnerabilities
TYPICAL ENGAGEMENT
ML PENTESTING
& SECURITY ASSESMENT