Cyberis Reply achieves CREST accreditation for Security Testing of AI

We’re pleased to announce that Cyberis Reply has achieved CREST accreditation for Security Testing of AI, building on our existing penetration testing accreditation. This provides independent assurance of our specialist capability to assess the security of generative AI and large language model-enabled systems.

For our team, this is an important milestone in the development of our AI security testing service. For our clients, it provides confidence that the expertise, methodology and quality controls behind that service have been assessed against an independent, AI-specific standard. CREST’s assessment considers practitioner competence, testing approaches, governance and the evidence supporting security conclusions.

What this means for our clients
Choosing an AI security testing provider requires confidence in their ability to understand your implementation and produce meaningful results. This accreditation gives clients an independently assessed benchmark to support that decision. It recognises our capability as a testing provider. 

Our focus as an accredited body is: helping clients understand what could go wrong, what we have demonstrated during testing and which improvements will reduce their exposure. Our reporting connects technical findings to business risk and provides remediation advice tailored to the client’s environment.

Our commitment to responsible AI

This accreditation sits alongside our commitment to the CREST AI Charter, which we signed earlier this year. The Charter is a voluntary public commitment to support CREST’s AI Principles, including accountability, transparency, human oversight, security and responsible data handling.

These commitments address complementary responsibilities: assessing the security of our clients’ AI systems and using AI responsibly within our own work.

Our approach is clear: AI supports the decision. It does not own it. Where AI assists our consultants, its outputs are reviewed and validated. Technical conclusions, risk assessments and the advice our clients rely on remain the responsibility of skilled professionals.

To discuss the security of an AI-enabled application, assistant or agentic workflow, speak to our team about an assessment tailored to your environment.