This work presents a safety-by-design assurance architecture combining model-level supervision, such as Scientist AI, with system-level controls over scaffolds and harnesses, independent verification, monitoring, and evidence infrastructure, supported by governance for accountability and evidence interoperability.
Abstract
Incidents show that AI safety failures often arise across multiple layers. We present a safety-by-design assurance architecture combining model-level supervision, such as Scientist AI, with system-level controls over scaffolds and harnesses, independent verification, monitoring, and evidence infrastructure, supported by governance for accountability and evidence interoperability.
This seminar explores how to make AI-assisted academic research verifiable, meaning inspectable, reproducible, and defensible by turning vague prompting into bounded, auditable workflows. It offers practical methods for building audit trails, testing workflow reliability, and documenting AI use in ways that support pub...
This “Practitioner’s Digest” column highlights emerging software engineering for AI research presented at the IEEE/ACM International Conference on Software Engineering 2026, focusing on reliability, runtime safety, sustainability, requirements alignment, and scalable AI operations.
S. Abrahão, Miroslaw Staron, S. Abrahão et al.· IEEE Software· 0 citations
It is argued that agents lower the cost of maintaining tests, commit histories, repository structure, instructions, and decision records while making their benefits immediate while making their benefits immediate.
Problem. The rapid deployment of large language models (LLMs) into safety-critical roles, including mental health support, education, and professional guidance, has created a gap between the safety claims AI developers publish and the behaviors their systems exhibit in practice. Current documentation practices, primari...
Kalvin Thuan-phong Khuu· Proceedings of the ACM/IEEE...· 0 citations