The presentation will discuss aspects of trust, security and safety of current AI systems and give some basic steps & principles to improve these factors. We shed light on the three main roles of AI in security, i.e., as a defence, as a tool for adversaries, and as a vulnerability in itself. Key takeaways:

 

- What are the main issues?

- How can we achieve security/safety/trust by design/default?

- Minimum requirements for compliance vs. Minimum requirements for peace of mind.