Monika Adamczyk

 Artificial Intelligence finds practical use in every aspect of our life and concrete implementations of innovative AI solutions can be found in almost every sector of the economy. AI systems can help increase productivity and effectiveness of our work or simply deal with mundane tasks. There are clear benefits stemming from use of AI, but as with any other technology, there are associated security and privacy risks. Trustworthy AI must be designed and implemented in such way that while they provide a satisfactory level of protection, they do not impact their intended functionality. 

ENISA has been focusing for the last few years on cybersecurity and privacy of AI and has developed multiple guidelines in this area, including an AI threat landscape, security controls for ML and an assessment of security and privacy context in specific AI use cases.

During the proposed talk, I will describe the AI threat landscape, how to choose security and privacy controls for AI systems and the possible trade-offs between cybersecurity, privacy and functionality.