Rishub Jain
CEO, Research Engineer
Sampura Research | Ex-DeepMind
Registration open
Warsaw AI
Safety Day
One day for the people shaping safer artificial intelligence—across research, policy and practice.
24 October 2026Faculty of Psychology, University of Warsaw · Stefana Banacha 2D, Warsaw
Speakers
Chloé Touzet
Policy Lead
SaferAI
Risto Uuk
Head of European Policy and Research
Future of Life Institute
Karolina Korgul
Doctoral Researcher
Google DeepMind | Oxford Internet Institute
David Quarel
Head of Teaching
ARENA
Christopher Stamper
Biosecurity Researcher
SecureBio AI
Jakub Kryś
Research Scientist
SaferAI, ERA
David Williams King
AI Safety Research Manager
ERA Cambridge
Jan Betley
Researcher
Truthful AI
Anna Sztyber-Betley
Assistant Professor
Warsaw University of Technology | Truthful AI | AI Safety Poland
Justin Olive
CEO
Generality Labs
Wojciech Dzięgiel
Board Member
Casimir Pulaski Foundation
Michał Kubiak
AI Policy Expert
AI Safety Poland
Michał Ołowski
AI Researcher
NASK
Chris Szulc
Managing Director
EA Poland
Jakub K. Nowak
Executive Director
AI Safety Poland
Maciej Chrabąszcz
AI Safety Researcher
NASK
Jan Broniowski
LLM Consultant | AI Security Expert
Sandra Sobańska
Frontier Lab Employee Survey
Włodzisław Duch
Nicolaus Copernicus University in Toruń
Szymon Łukasik
Director of AI Safety Research Center
NASK
Psyho
Who is it for
Curious about AI safety or governance and looking for a way in
Already in the field and seeking collaborators, mentors or a next role
Recruiting, researching or looking for institutional partners
Based in Poland or the wider region and committed to reducing advanced AI risks
Programme
Agenda
Two parallel tracks, brought together by shared keynotes, lunch and the film premiere.
Doors open09:00
Coffee & registration before the programme begins at 10:00.
- 11:10
Everyone together
Shared programme from 10:00 to 11:10
Opening & welcome
Sampura Research: Hill-climbing Alignment using Humans
Rishub JainCoffee break
- 12:4011:10–12:40
Choose one track
Parallel sessions
Track 01 Technical AI Safety
Frontier LLMs RL training and its failure
David QuarelValue Leakage: An LLM's Answers Are Silently Shaped by Its Own Values
Jan Betley · Anna Sztyber-BetleyHow to choose the right problems in AI safety
David Williams King
Track 02 AI Governance
From Rule-Takers to Rule-Shapers: AI Governance from the perspective of Middle Powers
Michał KubiakTwo kinds of "loss of control": Frontier Risk and European Sovereignty
Chloé TouzetDeepfakes and the Future of Trust: Why AI Governance Matters
Michał Ołowski
- 15:05
Everyone together
Shared programme from 12:40 to 15:05
Lunch & networking
AI safety film premiere
Enforcement Has Begun: the EU AI Act and Systemic Risks
Risto UukShort break
- 16:3515:05–16:35
Choose one track
Parallel sessions
Track 01 Technical AI Safety
You could probably make a better alignment eval if you tried
Justin OliveBiosecurity in the Age of AI
Christopher StamperShort introduction to Robotics Safety
Karolina Korgul
Track 02 AI Governance
Are existential risks a real problem right now?
Jakub K. Nowak · Chris SzulcGuarding the NATO's Eastern Frontiers – Defense AI in Poland
Wojciech DzięgielIntroduction to AI Verification
Jakub Kryś
- 17:00
Everyone together
Shared programme from 16:35 to 17:00
Coffee break
- 18:0017:00–18:00
Choose one track
Parallel sessions
Track 01 Technical AI Safety
Follow the Data: Understanding the AI Systems Attack Surface
Jan BroniowskiAI Monitoring Must Shift Beyond the Chain-of-Thought
Maciej Chrabąszcz
Track 02 AI Governance
Mniej oczywiste ryzyka rozwoju AI
Psyho · Włodzisław Duch · Szymon ŁukasikModerated by Sandra Sobańska
The 1:1 layer
Make the meeting you came for.
Inspired by the meeting culture at EAG events, protected gaps in the programme turn introductions into focused conversations.
- 01Create a profile around what you work on and what you need.
- 02Discover relevant attendees in Swapcard.
- 03Send a meeting request with a short reason.
- 04Meet in a dedicated slot during the day.
15:30
Your next collaborator20 min · Meeting point C
A lightweight way to find collaborators, mentors, peers—or your next role.
Powered by
AI safety, on the big screen.
The first Polish-language popular-science video explaining what AI safety is—with English subtitles.
- Produced by
- Piotr Kędziora
- Runtime
- 20–25 minutes
- Subtitles
- English
Made possible by a Rapid Grant from BlueDot Impact.
One ticket, the whole day
99 PLNIf the fee would be a barrier, email us for a complimentary ticket.
Included
- Vegan lunch
- Coffee service all day
- Swapcard 1:1 meetings
Registration always has to be confirmed—signing up does not guarantee a place. Your answers in the form decide whether you are admitted.
Reserve your placeBuilt together
Partners, and what each one does.
Organising & strategic
Media partners
Media partner
NASKPolish research and development organisation in digitisation and cybersecurity.Media partner
TrajektorieScience communication about the long-term future.Media partner
Center for Artificial IntelligenceResearch center for advanced artificial intelligence.Media partner
Wroclaw AI TeamWrocław-based AI community.
NASKPolish research and development organisation in digitisation and cybersecurity.Media partnerTrajektorieScience communication about the long-term future.Media partnerCenter for Artificial IntelligenceResearch center for advanced artificial intelligence.Media partnerWroclaw AI TeamWrocław-based AI community.Good to know
FAQ
How much is a ticket?
Tickets are 99 PLN. Every ticket includes the full programme, a vegan lunch and coffee all day.
Does registering guarantee a place?
No. Registration always has to be confirmed by our team—signing up does not guarantee a spot. Your answers in the form decide whether you are admitted, and we will email you once your place is confirmed.
What does the ticket include?
Admission to the full programme, a vegan catered lunch, and coffee service throughout the day.
What language is the event in?
The conference is held in English, start to finish. The Polish-language film premiere includes English subtitles. The panel closing the conference will be held in Polish.
24 October 2026 · Warsaw
Your next conversation could change your direction.
AGI?
We found it. It asked for good governance.