Foren

AI Forum 2023

zum Kalender hinzufügen

Freitag

10.

November 2023

AI Forum 2023

Ort: Forum Digitale Technologien

zum Kalender hinzufügen

Als gemeinsame Veranstaltung von Bundesamt für Sicherheit in der Informationstechnik (BSI), TÜV-Verband und Fraunhofer Heinrich-Hertz-Institut (Fraunhofer HHI) fand auch dieses Jahr wieder das AI Forum 2023 statt. In diesem Forum suchten wir mit internationalen Expert:innen und allen Interessierten den Dialog zum Stand und den Herausforderungen bei der Auditierbarkeit von KI-Systemen.

Im nunmehr 4. Internationalen Workshop 'Towards Auditable AI Systems: New Challenges introduced by Generative AI' lagt unser Schwerpunkt dieses Jahr auf der Prüfbarkeit von generativer KI (engl. 'foundational models'). Neben spannenden internationalen, interdisziplinären und inter-sektoralen Einblicken baut das Forum (das am Folgetag der TÜV AI Conference 2023 stattfand) auch auf der gemeinsamen Whitepaper-Serie der beteiligten Partner auf: zuletzt im Mai 2022 Whitepaper „Towards Auditable AI-Systems: From Principles to Practice“, in dem erstmalig eine „Certification Readiness Matrix“ (CRM) vorgestellt wurde.

Downloadbereich

Wojciech Samek, Fraunhofer HHI – Explainability 2.0 in the era of generative AI

Patrick Schramowski, DFKI – Can there be value in reflecting the world’s ugliness? Inappropriate and biased behaviour of generative models

Anna Wilhelm / Britta Sennewald, BSI – Large Language Models: Opportunities and Risks

Andreas Engel, Heidelberg University – Best Practices for developing and securing AI systems: a legal perspective

Alessandro Petri, QuantPi GmbH

Philipp Slusallek, DFKI – Trusted AI & CERTAIN

Marc Hauer, TÜV AI Lab GmbH – Using Assurance Cases to assure the fulfilment of non-functional requirements of AI-based systems

Bernhard Nessler, University Linz – Large Language Models and Certfication of AI/ML

Das AI Forum 2023 war eine gemeinsame Veranstaltung von

Ab 09:00 Uhr

Welcome Note

BSI | TÜV AI Lab | Fraunhofer HHI

Prof. Dr. Wojciech Samek | Fraunhofer HHI

Dr. Arndt von Twickel | BSI

Franziska Weindauer | TÜV AI Lab

Ab 09:20 Uhr

Session 1: New risks and chal-lenges introduced by generative AI

Prof. Dr. Wojciech Samek | Fraunhofer HHI

"Explainability 2.0 in the era of Generative AI"

Patrick Schramowski | DFKI & TU Darmstadt

"Can there be value in reflecting the world's ugliness?"

Anna Wilhelm / Britta Sennewald | BSI

"Large Language Models: Opportunities and Risks"

Moderation: Franziska Weindauer, TÜV AI Lab

Ab 10:20 Uhr

Coffee Break

Ab 10:40 Uhr

Session 2: Best-practices for developing and securing AI systems

Dr. Andreas Engel, LL.M. | Heidelberg University

"Best practices for developing and securing AI systems: a legal perspective"

Alessandro Petri | QuantPi GmbH

"From DIN SPEC to Reality: a German AI-Testing Engine tackles scalable AI Governance"

Prof. Dr. Philipp Slusallek | DFKI, Saarland University

"Trusted AI: The need for providing guarantees as the basis for future AI systems"

Marc Hauer | TÜV AI.Lab

"Using Assurance Cases to assure the fulfillment of extra-functional requirements of AI-based systems - Lessons learned"

Moderation: Prof. Dr. Samek Wojciech, Fraunhofer HHI

Ab 12:00 Uhr

Lunch Break

Ab 13:00 Uhr

Session 3: Challenges for LLMs in Europe

Dr. Tatjana Evas/LL.M. | EU Commission, DGCNECT (remote)

"The EU AI Act - operationalizing trust and innovation'

Prof. Dr. Philipp Hacker, LL.M. | European University Viadrina

"Foundation model regulation"

Dr. Bernhard Nessler | Johannes Kepler University Linz, Austria

Moderation: Rainer Plaga, BSI

Ab 14:00 Uhr

Panel 1: Technical & legal foundation, best-practices

Tatjana Evas, PhD/LL.M. | EU Commission, DGCNECT (remote)

Prof. Dr. Philipp Hacker, LL.M. | European University Viadrina

Dr. Daniel Weimer | Managing Director CertAI GmbH

Moderation: Dr. Arndt von Twickel, BSI

Ab 14:45 Uhr

Coffee Break

Ab 15:15 Uhr

Panel 2: AI Quality and Certification

Jörg Martin | Physikalisch-technische Bundesanstalt

Dr. Clara Neppel | Institute of Electrical and Electronics Engineers

Dr. Henrik Putzer | cogitron GmbH

Moderation: Prof. Dr. Wojciech Samek, Fraunhofer HHI

Ab 16:00 Uhr

Wrap-up What have we learned? What lies ahead?

TAG 1 - 10.11.2023, Freitag

Ab 09:00 Uhr

Welcome Note

BSI | TÜV AI Lab | Fraunhofer HHI

Prof. Dr. Wojciech Samek | Fraunhofer HHI

Dr. Arndt von Twickel | BSI

Franziska Weindauer | TÜV AI Lab

Ab 09:20 Uhr

Session 1: New risks and chal-lenges introduced by generative AI

Prof. Dr. Wojciech Samek | Fraunhofer HHI

"Explainability 2.0 in the era of Generative AI"

Patrick Schramowski | DFKI & TU Darmstadt

"Can there be value in reflecting the world's ugliness?"

Anna Wilhelm / Britta Sennewald | BSI

"Large Language Models: Opportunities and Risks"

Moderation: Franziska Weindauer, TÜV AI Lab

Ab 10:20 Uhr

Coffee Break

Ab 10:40 Uhr

Session 2: Best-practices for developing and securing AI systems

Dr. Andreas Engel, LL.M. | Heidelberg University

"Best practices for developing and securing AI systems: a legal perspective"

Alessandro Petri | QuantPi GmbH

"From DIN SPEC to Reality: a German AI-Testing Engine tackles scalable AI Governance"

Prof. Dr. Philipp Slusallek | DFKI, Saarland University

"Trusted AI: The need for providing guarantees as the basis for future AI systems"

Marc Hauer | TÜV AI.Lab

"Using Assurance Cases to assure the fulfillment of extra-functional requirements of AI-based systems - Lessons learned"

Moderation: Prof. Dr. Samek Wojciech, Fraunhofer HHI

Ab 12:00 Uhr

Lunch Break

Ab 13:00 Uhr

Session 3: Challenges for LLMs in Europe

Dr. Tatjana Evas/LL.M. | EU Commission, DGCNECT (remote)

"The EU AI Act - operationalizing trust and innovation'

Prof. Dr. Philipp Hacker, LL.M. | European University Viadrina

"Foundation model regulation"

Dr. Bernhard Nessler | Johannes Kepler University Linz, Austria

Moderation: Rainer Plaga, BSI

Ab 14:00 Uhr

Panel 1: Technical & legal foundation, best-practices

Tatjana Evas, PhD/LL.M. | EU Commission, DGCNECT (remote)

Prof. Dr. Philipp Hacker, LL.M. | European University Viadrina

Dr. Daniel Weimer | Managing Director CertAI GmbH

Moderation: Dr. Arndt von Twickel, BSI

Ab 14:45 Uhr

Coffee Break

Ab 15:15 Uhr

Panel 2: AI Quality and Certification

Jörg Martin | Physikalisch-technische Bundesanstalt

Dr. Clara Neppel | Institute of Electrical and Electronics Engineers

Dr. Henrik Putzer | cogitron GmbH

Moderation: Prof. Dr. Wojciech Samek, Fraunhofer HHI

Ab 16:00 Uhr

Wrap-up What have we learned? What lies ahead?

Druckversion

Haben Sie Fragen zu diesem Event?

Dr. Patrick Gilroy

Referent Künstliche Intelligenz und Bildung

+49 30 760095-360

patrick.gilroy@tuev-verband.de

Maria Sürig

Head of Events

+49 30 760095-330

maria.suerig@tuev-verband.de