Artificial intelligence (AI) has been a hot topic in recent years. Many previously impossible things have become possible with the use of AI. People are interested in the benefits of AI and different organizations are looking for ways to utilize it. However, alongside the benefits obtained, it should be remembered that AI systems also contain vulnerabilities that expose them to cyber attacks and thus information security risks. Consequently, the information security of these systems should also be considered in the same way as ordinary information systems.
Most of the information security risks AI systems contain are the same as the information security risks of traditional systems, but AI systems are also associated with risks specific to them. These risks are related to the data and models used by AI systems. Attacks on AI systems have been studied diligently in recent years from different perspectives, and based on the studies, several ways to secure these systems have been presented. However, the developed security measures are still incomplete. Consequently, organizations should be aware of the information security risks of AI systems when considering the implementation of the system in their own operations.
Marjo Hanhikoski
Centria University of Applied Sciences
Project assistant