Generative AI has spread to several fields of application. As simple as it sounds, it all relies on probabilities of a sequence occurring, based the same idea as the one behind an autocomplete, and input interpretations.

To close the gap between animal capabilities and robots’, we need physically intelligent bodies. Prof. Hughes’ lab is investigating on sensors and movement capabilities in soft robots. However, they need something more to get it all together... They needed AI.

Prof. Süsstrunk believes #education will prevent the spread of misinformation. Prof. de Rassenfosse adds that it has already changed to adapt around computers. Our goal is to teach children how to make the most out of the technology and not to blindly believe everything online.

How should legislative entities act so that they make sure such tools do not end up in ill-intended hands? Politicians already have set up rules for that, such as the EU AI Act. However, we don’t have enough ressource to enforce these #regulations.

What comes to mind when you hear "AI for healthcare"? Most likely disease detection (e.g. using X-rays), better treatment plans and finding new drugs. However, is AI for healthcare equivalent to curing a disease with AI?

Today’s round table animated by Prof. @marcelsalathe is gathering Prof.@ssusstrunk, Prof.@carmelatroncoso and Prof.@gderasse to talk about the potential risks to society of Generative AI. #AMLDGenAI23

A big thank you to Sir. Tudorache for this insightful talk!

One of the question that emerges is: « How will our children learn with these new AI-based tools ? ». They will eventually change the way our institutions work at large scale.

There are threats that could emerge from AI if not taken seriously. For instance, today, human does not have the monopole of the creation: IA occupies a big part of it and we need to understand its psychological effects on society.

Regulation will happen on three levels: 1. Developers of foundation models need to be transparent on their products 2. Models need to comply with a minimum set of rules, even if they aim to have a general purpose 3. The model needs to be transparent on the copyrighted material

As AI gained more and more attention, the EU realized that we needed to regulate such technologies. To quote the speaker, "Foundation models are a different breed of AI".

Now, Sir @IoanDragosT of the @Europarl_EN will be joining us on Zoom to talk about Foundation models in the EU AI Act. He will help us have a broader look on the impact of AI on the european society.