 Cyber criminals are studying up on artificial intelligence to use it for their advantage and weaponize it. At IBM Research we developed Deep Locker to demonstrate how existing AI technologies can be combined with malware techniques already being seen in the wild to create a new breed of attacks. Deep Locker is a new class of highly targeted and evasive malware. It's using AI to hide its malicious nature in popular applications until it reaches a specific target. It's identifying its target by means of AI models through indicators such as spatial recognition, voice recognition, or geolocation. What is unique about Deep Locker is that the use of AI makes it nearly impossible for security experts to detect the target ahead of time. The security community needs to prepare to face new levels of attacks, AI-powered attacks. We can't simply wait for these attacks to show up in the wild before we ramp up our defenses. IBM is leading the way to navigate the threats that AI-powered attacks could bring to organizations and help them to stay ahead of the cyber criminals.