Sci-fi often portrays artificial intelligence (AI) like this: a computer watches people for a while, blinks darkly and decides the solution to the world’s problems is to kill off the human race. Thankfully we are far away from that. But what AI is capable of today is simulating a specific human brain function – such as pattern recognition. And that’s very exciting for security.
AI makes security practical in the open world
The world is now open, causing disruption in many industries and changing the demands on security.
Pay television has evolved from a walled garden set-top box model to include online services. This opens up tremendous piracy threats that are a nightmare to control.
E-commerce and banking are undergoing open API mandates to enable information sharing among parties. This challenges the financial industry to grapple with cybercrimes such as man-in-the-middle attacks that hijack transactions and steal credentials.
For the automotive industry, cars are getting “smarter” with many connection points that can be exploited by hackers to tamper with systems or steal cars.
Security is no longer about closing borders and locking assets down. Security in the open world depends on being able to quickly spot something bad is happening or has happened. Such as someone putting a pirated video online, or trying to penetrate a service. As more pirated content goes online and more hackers go after high value targets, this means there is more data to collect and process to detect threats. A HUGE amount of data, in fact. You just can’t hire enough people to do this fast enough or cheaply enough.
AI enables us to adapt security to the demands of the open world. We can teach computers to recognize complex patterns among a sea of data that is growing at an exponential rate. Because AI is also advancing at an exponential rate.
AI gets exponentially better year on year
AI progresses so quickly because:
- Open source enables much faster Information sharing
- (Cloud) computing power follows the Moore’s law of exponential growth
- Learnings built on one another leads to better algorithms
Many applications are using AI to deliver great results faster and cheaper – image processing, real time translation and speech recognition to name a few. Their success is based on training computers to process a huge amount of data in order to recognize patterns and derive meaning from the patterns.
Promise: objects may be closer than they appear
When can we use AI to find pirated videos, even when they are modified to evade watermark or fingerprint detection? How can banks and payment service providers extend the use of AI to further reduce risk and keep pace with evolving cybercrime?
When can we take a nap on the way to the ski resort in our self-driving cars and arrive refreshed? 2 years? 10 years? Who knows, but I bet it’s sooner than we think. Our research team is exploring AI’s potential for security, and I for one find it full of promise.