Trustworthy & Ethical AI Conference

Patrick Van Eecke will discuss with you the impact of AI on our business practices and how European policy makers are planning to regulate the development and use of human-centric AI. In his talk, Patrick focuses on how Europe intends to regulate bias, fairness, ethics, trust and explainability of AI.

Patrick Van Eecke is Partner at global law firm DLA Piper and Professor IT & Law. He is familiar with the EU initiatives on Trustworthy AI and the clash between Law and AI/Robotics. He will discuss liability, IP (intellectual property, copyright, patents), competition law, data protection, data ownership, data donorship, ethics, fairness, explainability, bias, the lack of legal frameworks, and what local and European governments are planning to do about this (euRobotics, Robolaw project, Digital Charter on Robotics, Code of Ethical Conduct, ...).

Check out more information on the speaker Patrick Van Eecke

Issues of algorithmic fairness and transparency are relevant to both ethics and law. In this presentation, some of the ethical and legal challenges of AI will be discussed:

• Automated decision making: legal and ethical issues
• Why do algorithmic fairness and transparency matter?
• Algorithmic explainability as a legal obligation (GDPR)
• Bringing human in the loop: an ethical perspective
• Can AI help us make fairer decisions?
• Human-machine interactions

Check out more information on the speaker Nazanin Gifani

Vincent looks at the various AI initiatives at VDAB, its AI4Good strategy, the way applications are designed, and most of all, the way ethics, measurements through KPI's, explainability and fairness play a role in this. Vincent will explain how ethics-by-design works at VDAB.

Check out more information on the speaker Vincent Buekenhout

The past year, Brainjar has completed four similar document processing assignments where a 0% error rate was absolutely required. This presentation outlines how Brainjar used human-centric design and explainability to create machine learning systems that work together with humans to improve efficiency while reducing error rate. In this use case we will be talking about four cases specifically:

• Automatically processing vehicle rental orders via email or pdf
• Classifying scanned patient files
• Predicting unemployment benefits
• Processing scanned government documents

Check out more information on the speaker Deevid De Meyer

In this talk you will learn what Bias in ML models actually means. You will get insights in the complexity of the problem by going over of some examples and finally learn realistic ways to reduce bias whilst keeping the impact of those solutions on your application in mind.

Check out more information on the speaker Matthias Feys

Chatbots and conversational interfaces are taking over customer service departments by storm. In many companies, they provide first-line support to customers. Based on the Partena Ziekenfonds business case, Karel will share a few critical success factors:

• Make sure people know they are talking/typing to a bot
• What if privacy and authentication is needed
• The Partena Ziekenfonds business case
• Human take-over; first-line vs second-line customer service
• Explainability: looking into the Bot's Brain

Check out more information on the speaker Karel Kremer

Yves presents several instances where bias has posed a risk to the successful adoption of NLP systems, and discusses what techniques exist to discover these biases before the systems are put in production.

Check out more information on the speaker Yves Peirsman

Yves presents several instances where bias has posed a risk to the successful adoption of NLP systems, and discusses what techniques exist to discover these biases before the systems are put in production.

Check out more information on the speaker Yves Peirsman

• Ethical AI: are Belgian companies working on it?
• What are the difficulties in applying ethical AI
• Practical tools available for implementing AI Ethics
• Overview of AI initiatives with a focus on ethics:
• AI4Belgium
• Kenniscentrum Data & Maatschappij
• DigitalWallonia4AI
• EU High Level Expert Group

Check out more information on the speaker Jelle Hoedemaekers

Everybody is talking about AI. To turn this talk into action, organisations need to make sure that they use AI in an appropriate way. Firstly, they need to ensure that the developments are ethically sound and comply with regulations in all respects. Secondly, underpin the development by a robust foundation of end-to-end governance. Finally, support strong performance pillars addressing bias and fairness, interpretability and explainability, and robustness and security. That's what we've called "Responsible AI".
During the conference, we'll be looking at an example AI development process with focussing on the risks to be managed and the controls that can be established.

Check out more information on the speakers Martijn Cuypers and Hugo Pires

Paul explains how intelligence is added to a corresponding reality without introducing limitations into a world-of-interest. The outcome is obedience: a conflict with an obedient digital twin is a conflict with its real-world counterpart. An example in healthcare reveals how this augments human intelligence and enables intense interactions (among digital twins) without over-soliciting humans in terms of attention. A close and natural relationship of the obedient twins with a corresponding reality further facilitates explain-ability and data ownership. The absence of decisions carves out intelligence that eases ethical issues. Finally, the digital twins facilitate the explain-ability of any decision making AI that executes on a platform provided by these twins.

Check out more information on the speaker Paul Valckenaers