In January 2024 Microsoft chose Casaba to spend four months performing an in-depth security analysis of nine Copilot AI assistant implementations across the M365 product suite. Priority and focus was on the AI/LLM aspects specific to each Copilot.
Each Copilot was assessed in reference to the OWASP Top Ten for LLMs.
See the details of this public report to learn more about Casaba's AI/LLM security testing methods and deliverables.
Our advanced test automation for AI/LLM testing
Our testing techniques and methodology
Our report deliverables
We work at the forefront of Generative AI security. We are trusted by the world's leading AI service providers to fully vet security of the platforms and applications built on top. Our approach uses our own advanced test automation along with the latest testing techniques to ensure deep and thorough coverage of your AI platform and product security.
Learn MoreFor over two decades our specialty has been providing superior infrastructure and application security testing services to companies who already have a mature secure software development practice. We are called to validate the effectiveness of security controls in Cloud infrastructure and application code while also identifying the most critical and elusive exploits.
Learn MoreWe work closely with key stakeholders to create and manage robust cybersecurity strategies and programs that ensure the highest level of security, resilience, and compliance. Through strategic planning and tailored solutions, we help companies achieve comprehensive protection against cyber threats and maintain regulatory adherence.
Learn MoreOur experience with infrastructure and application security naturally lends itself to red team operations. Call us for your next attack simulation or classic pen testing red team op. Learn More
We have studied and tested IoT security from the cloud to the edge, including down to the firmware and hardware designs of embedded devices. Learn More
Still an important function of a shift-left practice, we can assist in building threat models and delivering actionable reports. Learn More
We have worked at the forefront of SDL for years, before most companies had even begun building their own secure development programs. Learn More