Cyber & IT Supervisory Forum - Additional Resources

ARTIFICIAL INTELLIGENCE AND CYBERSECURITY RESEARCH

Description:

Testbeds are required to study and optimise the performance of ML-based tools and technologies used for cybersecurity. Objectives:

1. Develop test beds to optimise the performance of AI/ML used in cybersecurity.

Entities:

• Security researchers • Application developers

Beneficiaries:

• Providers of AI tools and solutions

Existing research:

Much of the existing research effort is focused on achieving optimal accuracy in a simulated environment, which usually does not reflect the performance achieved under real-world conditions. Computational complexity and real-time operation must be taken into account, especially when the system being monitored is of critical importance. In this direction, efforts are needed to improve and construct test environments to study and optimise the performance of ML-based cybersecurity tools and technologies.

Standardised frameworks assessing the preservation of privacy and confidentiality Type: AI for Security Description:

Standardised frameworks assessing the preservation of privacy and confidentiality of the information flows as well as of the designed solutions need to be developed.

Objectives:

1. Preservation of privacy, e.g. training data and confidentiality of the information flows in systems so that the characteristics of the systems are not indirectly exposed and potentially classified information is not revealed; 2. Privacy preservation and confidentiality offered by the designed solutions.

Entities:

• Security researchers • Application developers • GDPR-related specialists

34