Cross-Sector Application of ISO/PAS 8800 for AI & ML Safety
Guidance on applying ISO/PAS 8800 across industries for AI/ML safety assurance.
Bridging the Gap Between IEC 61511 and Use of AI in Plant Safety
Bridging deterministic safety and AI uncertainty: a roadmap to reconcile IEC 61511 with the future of intelligent plant systems.
Uncovering Unsafe Feature Interaction in Vehicle Control using Generative AI
Harnessing LLMs and Digital Twins to uncover hidden ADAS risks faster, cheaper, and smarter.
A Fuzzy Logic Language for Trustworthy Confidence in Assurance Cases
Certus: Bringing clarity to assurance confidence with fuzzy logic and a language engineers can trust.
Not a “One Size Fits All” for AI in Safety Assurance
A call to tailor LLM evaluation methods to each assurance case use, risk profile, and role.
Balancing AI’s Promise and Peril in Assurance
A practical framework for weighing the risks and rewards of applying LLMs in safety assurance cases.
From Standard to System: ISO/PAS 8800 Assurance in Practice with Socrates
From design to deployment, ISO 8800 defines how to argue AI safety with logic, evidence and live metrics.
First of its Kind – The NRC Sees the Argument
First instance of the use of an assurance case in the US NRC regulatory environment, using structured argumentation.
LLM & Structured Argumentation Leads to Trust
LLMs support safety-critical decisions by evaluating structured arguments, identifying gaps, surfacing doubts, and generating defeaters to challenge bias and strengthen trust.
Establishing Trust in Data for Critical Decision Making
Using structured argumentation to establish trust in In-line Inspection (ILI) results used to evaluate the integrity of a pipeline.
Functional Safety Requirements for Artificial Intelligence and ML Systems
Discover how Metamorphic Relations offer a breakthrough in defining functional safety requirements for AI/ML, tackling the ‘black-box’ problem.
Confirmation Bias & Safety Management Systems
Discover how Eliminative Argumentation and AI enhance Safety Management Systems by reducing bias and providing real-time performance insights through KPI integration.
Electric Over Water – Analysis of Hazards for an All-Electric Floatplane
Electric propulsion is reshaping aviation, but can safety keep pace with innovation? Explore how Eliminative Argumentation addresses the hazards of tomorrow’s aircraft.
Utilizing an Assurance Case Argument to Drive Development
How can a safety case transcend being merely a final step in the safety process?
Adding Defeaters to Confidence Assessment
Take account of defeaters and doubt in confidence assessment methods, including those used in our Socrates product.
Assurance Cases in the Large
See how CERN’s Machine Protection System uses structured argumentation to mitigate risks in the world’s largest particle accelerator, ensuring operational safety.
Generating Defeaters with Gen AI
Use Generative AI to help brainstorm defeaters for structured arguments.
Challenging Autonomy with Combinatorial Testing
With trillions upon trillions of possibilities, how could you identify a set of test cases for an autonomous system that is both manageable and adequate?
Safety Integrity Levels for Artificial Intelligence
Re-thinking how level-of-rigour approaches for conventional software are applied to AI-enabled systems.
Incremental Assurance Through Eliminative Argumentation
How can an assurance case represent how confidence in an argument changes over time?
Assurance Case Development as Data
Could AI learn how to guide a safety engineer in the development of a safety case?
Patterns for Security Assurance Cases
Re-usable patterns, based on NIST 800-53, for creating security assurance arguments.
Safety for Cyber-Physical Systems
A practical approach to applying formal methods to guide the development of numerically intensive control systems.
Eliminative Argumentation for Arguing System Safety
Using the power of “doubt” to overcome confirmation bias and harden a safety assurance
case.
Bridging the Gap Between ISO 26262 and Machine Learning
Helping automakers and suppliers reconcile the use of ML with the conservative principles of functional safety.
Reducing the Feature Interaction Explosion Problem
Morse: a method and tool to simplify the Feature Interaction analysis using abstract subject matter knowledge.
Practical Uses of Formal Methods in Development of Airborne Software
Using formal methods to find what testing missed.
Creating Safety Assurance Cases for Rebreather Systems
An early example of safety assurance case development for a rebreather system per EN 14143, utilizing Goal Structuring Notation (GSN).
