Trustworthy AI regulations and their industrial/societal implications – a TECoSA Workshop with Digital Futures

All are welcome to this hybrid workshop to promote stakeholders & community dialogues on AI in industrial and societal applications, emphasizing ongoing regulatory work, gaps (regulators/research/industry) and innovation ecosystem creation.

For onsite participation and free lunch, please register here by 28 March.  Please note places are limited to 40, so first come first served.

Venue: Digital Futures hub, Osquars Backe 5, floor 2, KTH Campus, Stockholm  (Directions: https://www.digitalfutures.kth.se/contact/how-to-get-here/)
Or join us by Zoom:  https://kth-se.zoom.us/j/63920431339    (Meeting ID: 639 2043 1339)

WORKSHOP CHAIRS AND INITIATORS

ABSTRACT

Current rapid technological developments in Artificial Intelligence (AI) pose plenty of innovation opportunities and trustworthiness challenges – affecting all industrial domains and society!

The AI developments are taking place simultaneously with many technological innovations (e.g. sensors, communications, computing, additive manufacturing and augmented reality, to name a few) – leading to a socio-technical paradigm shift. This shift is reflected in changing regulations, standards, business models, insurance schemes, and new companies and collaborations. Specifically, in the digital decade of the EU, new regulations on AI, data, cyber-security, and digital services are being pursued, with related efforts in other parts of the world! At the same time, the legislation and standards are also evolving in virtually all industrial domains concerning AI and trustworthiness, some domains of which are already heavily regulated and some less so.

With all these developments, it is easy to get lost, and it is essential to understand that there will inherently be many (different) perspectives or viewpoints on what even the concepts of trustworthiness and AI mean, especially because of their broad applicability across industrial and societal domains. However, this challenge makes promoting dialogues and interactions between different stakeholders and communities even more important–representing this workshop’s important goal!

The following questions are of specific interest for this workshop:

• What is the state of the current AI regulation, and what are the next steps (regulations, guidelines, standards, etc.)
• What kind of regulations do we need? What are socio-technical considerations in balancing innovation and trustworthiness (cmp. for example, “AI alignment, An open letter to pause AI experiments”)?
• How do we reconcile the plethora of regulations and standards that will be overlapping and partly conflicting, reflecting different standpoints and inherent dependencies and tradeoffs (e.g. open data/learning/Explainable AI vs. privacy/confidentiality and IP; safety vs. performance/availability, etc.)?
• How does the AI act relate to and potentially conflict with other regulations, including domain-specific ones?
• What are the potential synergies? When/how can we reach convergence?
• How can we better bridge the gaps between regulators, industrial practitioners and researchers?
• How do we create innovative ecosystems promoting trustworthy AI and other sophisticated autonomous systems?

PROGRAMME:

[click here for bios, titles and abstracts of presenters and panelists]

12:00 – Informal icebreaker lunch

13:00 – Session 1: Keynote and Invited short talks (1h+30 mins)

• Keynote talk by Tatjana Evas (EC) (30 mins, 20+10)
• Short talks to set the context for the industrial panel (10+5)!
  • Fredrik Heintz, Professor LIU (AI)
  • Eduardo Gill-Pedro, Associate professor, Lund University (law)
  • Anna Felländer, Anch AI
  • Hans Hedin, Swedish Post & Telecom Authority (PTS)

14:30 – Leg stretch, coffee, and change to break out groups (15 mins)

14:45 – Session 2: Break out groups and brief reporting

• 20 min. work session in groups, with one person providing a summary with 2 highlights: Main problems to be resolved? Most important action ahead? Note-taking in a shared doc. (verbose is absolutely fine as long as 2 highlights are generated!)
• 20 min: Brief summaries per group (1 slide per group)

15:25 – Session 3: Panel debate – industrial responses (40 mins)

• • Shiva Sander-Tavallaey (ABB)
  • Maria Ramstedt (Ericsson)
  • Luis Martinez (Volvo cars)
  • Anna Sööder (Schibsted)
  • Potential further industrial representative

16:05 – Next steps and Wrap up (10 mins)

16:15 – End of Day