Sg.hu – Cyber ​​attacks – artificial intelligence can play an increasing role

Criminals are increasingly relying on algorithms.

Mikko Hyppönen, Research Manager, F-Secure IT Security Company said, anticipates that in the future, artificial intelligence will play an increasing role in cyber-actions by criminals. The expert said that malware developed and used by attackers will soon be automatically modified by algorithms so that they are not detected by anti-virus software.

Even today, cybercriminals regularly check whether or not Trojans used for extortion are detected by anti-virus software and change them if necessary. This task can be taken over by machine learning in a short time. Offenders may then be able to modify the applications used to execute the actions every 15 seconds to make it more difficult to discover and filter out these solutions. The only solution could be for security companies to change their systems in the same way. In this case, a competition between the artificial intelligences applied on the good and the bad side would start immediately.

Hyppönen, however, drew attention to the fact that self-learning systems based on machine learning are so complex and opaque that people can barely understand how they work. That’s why it doesn’t feel good at all to build solutions that no expert himself understands. At the same time, this feeling will become stronger as more artificial intelligence is involved in the fight against cyber-attacks.

The director of research at F-Secure called the undiscovered vulnerabilities in older WLAN protocols a potentially major threat. According to the manager, the nightmare-like scenario would be for someone to find a flaw and then infect all computers in an office or all smartphones in the subway with malware. No one has seen anything like this before, but it could happen at any time.

.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.