Fighting AI with AI: Making Safety-Critical Systems Trustworthy

Kari Jaaskelainen

26 Nov 2025 — 1 min read

When AI helps fly aircraft or drive cars, “just trust the model” isn’t enough. Deep neural networks are opaque, and the gap between plain‑English requirements and low‑level code blocks traditional safety assurance.

This paper proposes using foundation models to assure AI itself, via two complementary tools:

REACT: Uses Large Language Models to translate natural-language requirements into precise, checkable specifications. It flags ambiguities, suggests refinements, and can generate tests—enabling earlier verification and validation.
SemaLens: Uses Vision-Language Models to reason about, test, and monitor DNN perception with human-understandable concepts. It helps surface failure modes and corner cases that matter in the real world.

Together, they form a pipeline from informal requirements to validated implementations for domains like aerospace and autonomous vehicles—bringing explainability, consistency, and continuous monitoring to AI-enabled, safety-critical systems.

By Anastasia Mavridou, Divya Gopinath, and Corina S. Păsăreanu. Read more: https://arxiv.org/abs/2511.20627v1

Paper: https://arxiv.org/abs/2511.20627v1

Register: https://www.AiFeta.com

#AI #SafetyCritical #Assurance #LLM #VLM #AutonomousVehicles #Aerospace #DeepLearning #TrustworthyAI #RequirementsEngineering

Meet GRACE: a moral governor for safer, more transparent AI

AI agents are getting powerful—so how do we make sure they do the right thing, not just the effective thing? Meet GRACE, a reason-based moral governor that keeps AI behavior aligned with human norms by separating moral reasoning from goal-driven decision-making. * Moral Module: uses deontic logic and explicit reasons

Why some fine-tuned LLMs miss phishing—and how to fix it

Not all fine-tuned LLMs spot phishing equally. A new study tests Llama 3.1 8B, Gemma 2 9B, and Mistral on high-stakes phishing detection—and uses SHAP and mechanistic interpretability to reveal why models do (or don’t) generalize. * Architecture × data diversity matters: Gemma 2 9B hits state-of-the-art performance (F1

AI that explains itself—by following the science

AI that explains itself—by following the science Black-box AI can be powerful, but it often can’t tell us why it made a decision. Concept Bottleneck Models (CBMs) try to fix that by predicting human-understandable concepts first, then the final answer. The catch: standard CBMs ignore domain-specific cause-and-effect and

Meet GenomAgent: A Team of AI Specialists for Smarter Genomics Q&A

TL;DR: Finding reliable answers in genomics is hard. GenomAgent turns one big AI into a coordinating team of specialists—and beats the current leader by 12% on a key benchmark. Genomic facts live across many databases. Standard chatbots struggle because they can’t flexibly query those sources. GeneGPT added