Les extendemos una cordial invitación a nuestra próxima charla virtual, centrada en la intersección entre la confianza social y la fiabilidad en la inteligencia artificial y la arquitectura de software.
Detalles del Evento:
- Título: Observability vs Explainability: Where Social Trust meets Trustworthiness
- Ponente: Dr. Hernán Astudillo (Universidad Andrés Bello, Chile)
- Fecha: Miércoles, 2 de septiembre de 2026
- Hora: 15:00 a 16:00 (GMT-3)
- Conexión ZOOM: https://salavirtual-udelar.zoom.us/j/83002162564
Abstract: The growing use of agentic AI is feeding demand for “trustworthiness”. Ironically, as Han remarked, this denotes lack of trust, in the normal inter-agent sense: we demand transparency (or other forms of predictability) because we have no trust on the agents behavior, and this transparency makes their guarantees credible. Explainability was born as a key concept in AI, but is colonizing other fields, including software architecture. Or so we thought. We found in recent field work with collaborators that, while researchers are keen on Explainability (a property of computations), practitioners are keen on Observability (a property of systems). Also, recent work in other fields is shifting the focus from trust as a reflection of partner attributes to trust as an internally calibrated judgement shaped by contextual contingencies. This talk will explore some implications of this shift for software architecture, and will describe our ongoing work on the topic.
Mini-Bio del Ponente: Hernán Astudillo is a Ph.D. in Information and Computer Science (Georgia Tech, 1996) and a Professor at Universidad Andrés Bello (Chile), where he chairs the doctoral program on Advanced Systems Engineering. In earlier lives, he has been Senior-, Lead- and Chief Architect at MCI Systemhouse, Financial Systems Architects, and Solunegocios, and professor at Universidade de Sao Paulo (Brasil) and Universidad Técnica Federico Santa María (Chile). He does research, teaching and technology transfer on technological decision-making using imperfect information, focused on identification, evaluation and reuse of Software Architecture decisions, including tactics, patterns and trade-offs for Contemporary Software Systems, lately microservices and IoT. He is a board member of the Chilean Society of Computer Science (SCCC), and member of IEEE, ACM, SCCC, ALA, IASA, and CI-Chile. Perfil de investigación: https://researchers.unab.cl/
Esperamos contar con su valiosa participación.
