AI, the Cyber Criminal Next Door?



Jeremy Straub has written a very interesting article for The Conversation blog regarding the inevitability of AI assisted cyber-attacks, and how these attacks may in fact be more imminent that we think.

We have previously explored on Wave of Automation how AI may be applied to increase the effectiveness of cyber security by assigning large volume routine tasks to machines, and allowing human experts to concentrate their advanced decision making skills towards higher value tasks.

Straub discusses the other side of the exchange, where cyber criminals implement AI assistants to help launch more sophisticated attacks on unsuspecting victims.  An AI capable machine, with its ability to process vast amounts of collected data (through phishing or other means) may quickly uncover and connect hidden correlations or relationships that a human may use to identify a target.  When set up, an attack could then be launched from the same AI machine with incredible speed.  With such a methodology, smaller, more undetectable attacks may be conducted many times over with automation, allowing an attacker to steal or disrupt on a massive scale.

If cyber criminals can apply AI to their exploits, and cybersecurity experts can also apply AI to help thwart those criminals, one can easily see the spark of an AI arms race in cyber, with both criminals and defenders racing to develop the most advanced capabilities to exploit each other's vulnerabilities.

If this scenario is then carried out to its logical conclusion or end-game, we may see a world where advanced AI machines will handle all aspects of cyber crime, warfare and defense, self-evolving as required to gain supremacy.

For now, Straub feels we are still a long way from living in this type of world.  Our current machines are only capable of doing small, specialized AI tasks.  The machines themselves cannot attack us, a human being must ultimately make the decision whether to use AI technology to do good, or to do bad.

Read the full article here. 

What is your take? How do you feel that AI may affect your security in the digital world?

Comments