Lunch Talks "AI and Ethical Scientific Practices"
Event details
| Date | 26.05.2026 |
| Hour | 12:15 › 13:45 |
| Speaker | Orr Paradise; Gaia Barazzetti; Valentin Conrad; Pablo Diaz; Cécile Hardebolle |
| Location | |
| Category | Conferences - Seminars |
| Event Language | English |
Join the next Lunch Talks session, "AI and Ethical Scientific Practices", on May 26th, from 12:15 to 13:45, in BM 5202.
Language: English
Doors open from 12:05
Free lunch bag - water not included: bring your reusable bottle to reduce environmental impact
PROGRAM
Today's AI landscape can feel technologically inevitable: large models are used for everything, trained on everything, and trusted beyond what they can justify. In this talk, I will suggest that theoretical computer science offers a way to design and build alternatives. Drawing on work on Self-Proving Models and on research with the Cetacean Translation Initiative (Project CETI), I will describe how theory can help turn ideas such as verification and respect for the object of study into well-defined research questions.
Taking scientific integrity as a starting point, this presentation provides an overview of the legal and ethical rules that apply to AI research at EPFL. Topics include data protection, copyright, and ethical requirements for AI projects, particularly when people or personal data are involved.
As artificial intelligence becomes embedded in research practices, new ethical questions arise regarding how research is designed, conducted, and interpreted. This presentation focuses on the practical work of ethics committees in reviewing research projects involving AI. It highlights how committees examine assumptions, risks, impacts, and responsibilities throughout the research process, including the use of AI as a tool and as an object of study. The talk aims to clarify the specific contribution of ethics review in supporting responsible and reflective AI enabled research.
AI is increasingly present in research, whether as a tool or as an object of research, which raises specific risks if ethical considerations are overlooked. This talk presents a practical approach to identifying, assessing, and managing these risks throughout the research lifecycle by embedding responsible design practices from the start.
The event is free upon registration. Lunch bags will be offered to all registered participants. For sustainability reasons, plastic water bottles will not be provided. Please bring your own reusable bottle.
Language: English
Doors open from 12:05
Free lunch bag - water not included: bring your reusable bottle to reduce environmental impact
PROGRAM
- Talk 1: A Theoretical Lens on Responsible AI (10 min)
Today's AI landscape can feel technologically inevitable: large models are used for everything, trained on everything, and trusted beyond what they can justify. In this talk, I will suggest that theoretical computer science offers a way to design and build alternatives. Drawing on work on Self-Proving Models and on research with the Cetacean Translation Initiative (Project CETI), I will describe how theory can help turn ideas such as verification and respect for the object of study into well-defined research questions.
- Talk 2: Before You Train That Model: Legal and Ethical Basics for AI Research at EPFL (15 min)
Taking scientific integrity as a starting point, this presentation provides an overview of the legal and ethical rules that apply to AI research at EPFL. Topics include data protection, copyright, and ethical requirements for AI projects, particularly when people or personal data are involved.
- Talk 3: Ethics Committees and AI in Research: A Practical Perspective (10 min)
As artificial intelligence becomes embedded in research practices, new ethical questions arise regarding how research is designed, conducted, and interpreted. This presentation focuses on the practical work of ethics committees in reviewing research projects involving AI. It highlights how committees examine assumptions, risks, impacts, and responsibilities throughout the research process, including the use of AI as a tool and as an object of study. The talk aims to clarify the specific contribution of ethics review in supporting responsible and reflective AI enabled research.
- Talk 4: AI for research and research about AI: a responsible design approach (10 min)
AI is increasingly present in research, whether as a tool or as an object of research, which raises specific risks if ethical considerations are overlooked. This talk presents a practical approach to identifying, assessing, and managing these risks throughout the research lifecycle by embedding responsible design practices from the start.
- Panel for Q&A, open discussion
Links
Practical information
- General public
- Registration required
Organizer
- EPFL Library