We can expect that 2018 might turn pivotal to artificial intelligence cyber-attacks. New kind of hacking is knocking on the front door and we can do nothing about it. Artificial intelligence armed with complex algorithms is now a thing.
We can expect that 2018 might turn pivotal to artificial intelligence cyber-attacks. New kind of hacking is knocking on the front door and we can do nothing about it. Artificial intelligence armed with complex algorithms is now a thing. It is completely autonomous and is able to act on its discretion, delivering stealth, slow and untraceable blows. Machines taking on machines is not science fiction anymore.
Autonomous cyber-attacks are uncontrolled when once released. Like a virus, they can enter the enemy territory and stay there observing and evolving. Able to learn from the surroundings, AI may pose a serious threat to huge databases, personal computers, and whole networks as well. However, the speculative nature of AI prevents us from precise predictions.
Weaponized AI is already thrown on digital networks where it can stay for months or even years without being detected. With the help of AI, cyber-attackers can penetrate networks unnoticed and pervade the IT infrastructure. Progress in artificial intelligence is happening at such a fast pace that it will not be a surprise if in recent months such blending algorithms merge with the daily activity on the network. The more the attackers stay inside a network, the stronger they become. They learn the environment and gain knowledge about its security downfalls, users and might take over the system.
Comparisons to the HIV virus are not far-fetched. AI attacks will aim to compromise the defense mechanisms of a system and turn them against it. The firewall patterns and the working hours of security teams are easy to capture. Once the AI learns them it will manage to avoid them with an outstanding efficiency.
Imagine if AI takes control over AI assistants that do our schedule, place orders or manage our emails. The scenario may get even worse. What if AI observes your writing style so well it learns to mimic it? Sure, you use different language, depending on who you write but this data is easily obtainable.
Instead of fighting with ghosts, security teams should work on building AI able to fight counterfeit systems.