Giada Pistilli

@giadapistilli.com

2829 Followers 173 Following

Philosopher in tech, currently at Mistral AI. Doctor of talking machines, now teaching them good behavior.

My Spotify artist of the year: The Beatles
AIES 2024 poster of the paper "ELLIPS: An Ethics-Centered Approach to Research on LLM-Based Inference of Psychiatric Conditions"
Picture of Martinique
Interview "Il est important de démystifier cet outil"
Me giving the keynote for EuroSciPy 2023 at Basel University
Blue building with Swiss and French flags on it.
Me sitting in a panel with Monique Canto-Sperber and Alexei Grinbaum
Illustration of intersections between normative, prescriptive, and descriptive. Being normative, values expressed in an ethical charter inform both prescriptive (what uses of an ML artifact should be allowed or prohibited) and descriptive (what capabilities and possible failures need to be reported; dashed lines), while technical documentation of ML artifact’s behavior and capabilities inform what likely harms and possible rights violations need to be addressed in licenses and regulations (dotted line). Regulations also specify what technical information needs to be reported for AI systems, for example, in model cards (full line).
Table of Use cases: Ethical: How ought this technology be used? Legal: How shall this technology be used? Technical: How can this technology be used?
Screenshot of the Analytics India Magazine titled "The Pitfalls of Fear-Mongering in AI"