Safety First

About Safety First

Learn about the world of functional safety.

Publisher

Zachary Paul Lukasiewicz

Category

technology

Language

en

Explicit

No

First episode

30 Jul 2026

Latest episode

7 Oct 2026

Latest episodes

202 episodes in the feed.

1. 

   7 Oct 2026

   Safety and Artificial Intelligence for Road Vehicles: Dataset Design - Exercise

   In ISO eight eight hundred, the responsibility for dataset quality directly impacts system safety performance for automated road vehicles. When developing safety-critical automated functions, your training and validation datasets must be traceable, representative, and rigorously documented. Part of the Critical Systems Analysis functional-safety series on Safety and Artificial Intelligence for Road Vehicles. In this episode: This episode covers the foundational exercise in ISO eight eight hundred: designing your dataset strategy for safety-critical perception systems. Every dataset design begins with understanding your use case. Safety-critical datasets require multiple dimensions of specification. The data pipeline follows a disciplined sequence. Dataset quality ascends through layers. Your dataset control scope includes all samples you use for training or validation, their acquisition parameters, annotation provenance, and modification history. Reference: ISO 8800 is a Publicly Available Specification (PAS) addressing AI/ML safety in road vehicles; dataset design sits in the Architecture and Design phase, bridging AI safety objectives More in this series: IEC 61508 — Functional Safety Foundations (https://open.spotify.com/show/033VUtmeVDRN9FSv91ybsz) | SIL and Safety Integrity — IEC 61508 (https://open.spotify.com/show/0343BTpEo19OzjRotOEsZp) | Safety-Critical Software (IEC 61508-3) (https://open.spotify.com/show/0343WmXFWpDPZYJhW2ybtu) | Functional Safety Training and Certification (https://open.spotify.com/show/0347Nzok2mfgwvLEOxB0hK) | Functional Safety Assessment and Services (https://open.spotify.com/show/0343V0EvIKxsDAm2OuTdw3) | Functional Safety Compliance and Audits (https://open.spotify.com/show/0343UHM6sw4dfwZtYZ1pMH) Explore more from Critical Systems Analysis (https://www.criticalsystemsanalysis.com/). Partner with us (https://meetings.hubspot.com/benjamin-twombly/strategic-partnerships) Follow on LinkedIn (https://www.linkedin.com/company/criticalsystemsanalysis) Visit our website (https://www.criticalsystemsanalysis.com/) Read our feature on Automate.org (https://www.automate.org/robotics/news/critical-systems-analysis-advancing-functional-safety-and-systems-assurance-across-industries-critical-systems-analysis)

2. 

   7 Oct 2026

   Safety and Artificial Intelligence: General Workflow for Deriving Safety Requirements — Exercise

   Artificial intelligence is reshaping every industry, but with innovation comes responsibility. When AI systems make safety-critical decisions—from autonomous vehicles to medical diagnostics—the stakes demand rigorous engineering discipline. ISO eight thousand provides the systematic framework for deriving safety requirements that protect people and systems. Part of the Critical Systems Analysis functional-safety series on Safety and Artificial Intelligence. In this episode: We begin our journey into ISO eight thousand by understanding the scope of functional safety for artificial intelligence. At its core, functional safety means the system reliably performs its safety functions even when faults occur. The general workflow starts with understanding your context: who you're protecting, what the AI system does, and what the stakes are. Hazard identification is where you systematically ask: what could go wrong with our AI system? You examine failure modes, data quality issues, and edge cases. Once you identify hazards, assess risk using two dimensions: severity—how bad if the hazard occurs?—and probability of occurrence. Safety requirements transform general safety goals into specific, measurable, testable specifications. Reference: ISO/PAS 8800:2024 published December 2024; full title 'Road Vehicles — Safety and Artificial Intelligence' More in this series: IEC 61508 — Functional Safety Foundations (https://open.spotify.com/show/033VUtmeVDRN9FSv91ybsz) | SIL and Safety Integrity — IEC 61508 (https://open.spotify.com/show/0343BTpEo19OzjRotOEsZp) | Safety-Critical Software (IEC 61508-3) (https://open.spotify.com/show/0343WmXFWpDPZYJhW2ybtu) | Functional Safety Training and Certification (https://open.spotify.com/show/0347Nzok2mfgwvLEOxB0hK) | Functional Safety Assessment and Services (https://open.spotify.com/show/0343V0EvIKxsDAm2OuTdw3) | Functional Safety Compliance and Audits (https://open.spotify.com/show/0343UHM6sw4dfwZtYZ1pMH) Explore more from Critical Systems Analysis (https://www.criticalsystemsanalysis.com/). Partner with us (https://meetings.hubspot.com/benjamin-twombly/strategic-partnerships) Follow on LinkedIn (https://www.linkedin.com/company/criticalsystemsanalysis) Visit our website (https://www.criticalsystemsanalysis.com/) Read our feature on Automate.org (https://www.automate.org/robotics/news/critical-systems-analysis-advancing-functional-safety-and-systems-assurance-across-industries-critical-systems-analysis)

3. 

   6 Oct 2026

   Safety and Artificial Intelligence: Need for additional safety requirements on AI systems – Exercise

   Artificial intelligence systems introduce unique safety challenges that traditional engineering approaches may not fully address. When deploying AI in safety-critical applications, we must identify and mitigate risks emerging from the AI system's inherent properties: opacity in decision-making, potential for unexpected behaviors, and dependence on data quality. Part of the Critical Systems Analysis functional-safety series on Safety and Artificial Intelligence. In this episode: This exercise focuses on a critical gap in traditional safety standards when applied to AI systems. Artificial intelligence introduces behaviors that traditional safety standards do not account for. Safety risks in AI systems stem from multiple factors including data quality, model robustness, system integration, and human factors in operation. Once you have identified AI-specific risks, you must develop safety requirements to address each one. Safety requirements for AI systems span multiple categories: functional requirements ensure the system maintains safety, robustness requirements guarantee consistent behavior under varied conditions, and monitoring requirements enable detection of unsafe states. Verifying AI safety requires layered approaches that provide complementary assurance. Reference: ISO eight thousand eight hundred (Safety and Artificial Intelligence for Road Vehicles) is a technical specification (PAS status as of twenty twenty-four) extending ISO twenty-six More in this series: IEC 61508 — Functional Safety Foundations (https://open.spotify.com/show/033VUtmeVDRN9FSv91ybsz) | SIL and Safety Integrity — IEC 61508 (https://open.spotify.com/show/0343BTpEo19OzjRotOEsZp) | Safety-Critical Software (IEC 61508-3) (https://open.spotify.com/show/0343WmXFWpDPZYJhW2ybtu) | Functional Safety Training and Certification (https://open.spotify.com/show/0347Nzok2mfgwvLEOxB0hK) | Functional Safety Assessment and Services (https://open.spotify.com/show/0343V0EvIKxsDAm2OuTdw3) | Functional Safety Compliance and Audits (https://open.spotify.com/show/0343UHM6sw4dfwZtYZ1pMH) Explore more from Critical Systems Analysis (https://www.criticalsystemsanalysis.com/). Partner with us (https://meetings.hubspot.com/benjamin-twombly/strategic-partnerships) Follow on LinkedIn (https://www.linkedin.com/company/criticalsystemsanalysis) Visit our website (https://www.criticalsystemsanalysis.com/) Read our feature on Automate.org (https://www.automate.org/robotics/news/critical-systems-analysis-advancing-functional-safety-and-systems-assurance-across-industries-critical-systems-analysis)

Contact Safety First

Guest appearances

Does not typically book guests

Based on episode analysis; this does not confirm that the show is currently accepting guests.