The evolution from simple rule-based systems to complex, opaque models like LLMs necessitates a responsible approach due to their growing societal impact.
Responsible AI is defined by its benefit to all stakeholders, grounded in the core principles of fairness, transparency, and human-centered design.
Assessing an AI system's trustworthiness involves examining its training data, the scope of process automation, and how it uses powerful models like LLMs.
The EU AI Act aims to create a harmonized legal framework for trustworthy AI, preventing inconsistent practices and building consumer trust.
The legislation categorizes AI systems into four tiers—unacceptable, high, limited, and general purpose—each with distinct obligations for different actors in the value chain.
High-risk AI systems must adhere to strict requirements, including risk management, data governance, technical documentation, and human oversight.
AI systems must undergo a conformity assessment, often based on internal controls or future harmonized standards, to receive CE marking for market placement.
Non-compliance with the EU AI Act can result in substantial fines, reaching up to 35 million euros or 7% of global revenue for prohibited practices.
29:35Steffen Bosse, Mina Saidze, Ray Eitel-Porter & Björn Bringmann
28:05Toju Duke
25:51Dr. Marc Fuchs, Christoph Bräunlein, Eva Stepkes & Niklas Harzheim
22:44Olena Linnyk
31:15Thomas Wollmann, Gerrit Einhoff, Kara Sprague & Alexandra Wudel
30:28Alexandra Wudel, Madalina Florean & Laura Moritz
28:27PJ Hagerty
27:26Cassie Kozyrkov
Jobs that call for the skills explored in this talk.
TRUSTEQ GmbH
AIT AUSTRIAN INSTITUTE OF TECHNOLOGY GMBH