CERTIFAI Midterm Workshop: Advancing AI-Driven Certification

On February 13, 2025, Prague hosted the midterm workshop of the CERTIFAI project, bringing together key stakeholders, industry experts, and researchers to discuss AI-driven certification processes. The workshop provided a platform to review project milestones, explore use cases, and introduce new technological developments that enhance certification frameworks.

Setting the Stage for AI Certification

The event opened with a welcome session, followed by a ‘tour de table’ where participants introduced themselves. The agenda and the concept of the Advisory Board were explained, highlighting the role of stakeholders in guiding the project’s direction. The initial session provided an overview of CERTIFAI, detailing the project’s objectives, scope, timeline, and key contributors.

Addressing Certification Challenges

One of the core discussions revolved around defining the problems that CERTIFAI aims to solve. Legal uncertainties and the complexities of AI certification were highlighted, along with how CERTIFAI is working to mitigate these challenges through structured methodologies and AI-driven tools.

Exploring Key Use Cases

A series of presentations showcased how CERTIFAI is applied in different sectors:
Railway Certification: Addressing vulnerabilities and improving compliance processes.
Smart Grid Security: Enhancing cybersecurity and threat traceability.
Autonomous Shipping: Developing AI-driven certification mechanisms for maritime applications.
Fire Detection IoT: Strengthening security frameworks for IoT-based fire detection systems.

The Auditor’s Perspective on AI Certification

A dedicated session provided insights into how AI-driven certification preparation and execution are perceived from an auditor’s viewpoint. This discussion emphasized the need for transparency, consistency, and alignment with existing regulatory frameworks.

Technological Innovations in CERTIFAI

Several advanced tools and frameworks were introduced:
IEC Cybersecurity Standards Interpreter RAG: AI-based tool to assist in interpreting cybersecurity and compliance standards.
Scalable Formal Verification Framework: A new framework aimed at improving automated certification assessments.
CERTIFAI Integrated Framework: A tool designed to streamline certification applications.
Semi-Automated Threat Modelling: A demonstration of the platform’s ability to automate threat modeling processes, enhancing security evaluations.

Looking Ahead

The workshop concluded with a summary of key discussions and an invitation to stakeholders for continued collaboration. Participants were also given a pre-invitation to the final conference, where further advancements in CERTIFAI will be showcased.
By fostering collaboration between AI researchers, industry players, and regulatory bodies, CERTIFAI is paving the way for more efficient, transparent, and reliable AI certification processes. Stay tuned for more updates as the project moves towards its final phase!