How to protect artificial intelligence technology from hackers

UC Science Today

Artificial Intelligence has been developing fast – and it’s making more and more decisions on humans’ behalf. From simple web searches to e-commerce to self-driving cars. But researchers like Dawn Song, a computer scientist at the University of California, Berkeley, are struggling to find the best ways to protect this technology from hackers. “One thing we are looking at how we can use AI and deep learning techniques to develop stronger security capabilities to enable us to build better defenses. One of our recent works uses advanced deep learning techniques to try to identify software vulnerabilities in IOT devises.” Deep learning techniques are advanced algorithms, which allow researchers to process massive amounts of data. “The advantage of deep learning is that it can enable us to build very large data sets. And now we have very good computation framework and hardware to perform a variety of tasks.” But as machines becomes smarter, so do viruses that attack them, so scientists are challenged to stay one step ahead of hackers.