As one of the key technologies of the 21st century, artificial intelligence (AI) is rapidly emerging as a game changer in the economy and in society. AI is based on machine-learning algorithms that are run on computers and therefore, face security risks. What are the implications of current and future use of machine learning for cybersecurity and how is AI transforming information security?
While security firms and researchers have been using machine-learning models to better anticipate attacks and to accelerate the response to incidents, it is very likely that hackers are adopting the same technology to strike back. Machine-learning models have proven remarkably adept in crafting convincing fake messages, allowing malicious hackers to conduct more phishing attacks. This capability makes cybersecurity even more important for many machine-learning systems than for traditional computer systems. If, for instance, machine learning is used to detect spam and it matches words in a message that are only present in a dictionary, spammers can easily modify words to avoid detection. At the same time, machine learning can be poisoned by allowing the machine to learn false or wrong information, invalidating the primary purpose of the machine processes, namely feeding algorithms with incorrect data.
Against this background, this high-level event brings together cybersecurity and AI experts from diverse fields to explore the implications of these dynamics and to consider ways to increase awareness of and find solutions to these security risks.
Participation in this event is exceptionally free of charge. A light lunch will be provide from 12.30