Nearly all available public data is now available in foundation models. On Hugging Face there are over 900’000 models you can choose from. This clearly leads to a future of open and transparent artificial intelligence (AI) we’re already using and going to use in our business workflows. This is also confirmed by the recent Swiss Open Source Study where it has been reported that 40% of companies are using already Open Source AI tools like TensorFlow or Pytorch or open models like Llama, Mistral or Granite. With the increasing possibilities using AI, we have seen challenges when using public data, including problems related to quality, bias, data origin and trustworthiness.

In this afternoon event, we’ll provide insights into the art of the possible with open source AI starting with a point of view from Hugging Face and the journey how IBM created and open-sourced the IBM Granite model family. We’ll also touch legal considerations when using AI in context to the EU AI Act or the upcoming Swiss regulation of AI.

Schedule:

14:30h Welcoming of the participants

15:00h Opening notes
by Prof. Dr. Matthias Stürmer (CH Open & BFH) and Markus Danhel, Principal Sales Leader (IBM)

15:15h Distributing a Million Open Models in the Wild
Lewis Tunstall, Machine Learning Engineer at Hugging Face

15:40h Open Source Document AI
Peter Staar, Principal Research Staff Member, Master Inventor, Manager of "AI for Knowledge" group

16:05h Coffee Break

16:30h Open-sourcing in Artificial Intelligence
Siddhartha Singh, researcher at Bern University of Applied Sciences (BFH)

17:05h Legal issues with Open Source AI
Simon Schlauri, Attorney at Law, Ronzani Schlauri Attorneys

17:30h Closing notes
by Prof. Dr. Matthias Stürmer (CH Open & BFH) and Markus Danhel, Principal Sales Leader (IBM)

17:45h Apéro

To participate at the event please register below. If you have any questions to the event, please contact Kateryna Schütz, Director General CH Open.