Episode 54 Cybersecurity Challenges in the Era of Large Language Models

Tech Uncensored - An Altitude Accelerator Podcast

With the emergence of Large Language Models (LLMs) and the speed of adoption, how we communicate, how we research, how we improve our productivity has greatly transformed society. While these powerful tools offer remarkable capabilities in natural language understanding and generation of more accurate human-like language, they have also introduced new challenges and risks to the practice of cybersecurity and its adjacency into data privacy.

Data breaches and attacks have been prevalent in narrow AI but with increased vulnerabilities introduced by LLMs, we’re now seeing more sophisticated phishing attacks, manipulation of online content, and exploitation of privacy controls.

On April 22nd the Biden Administration signed Section 702 of FISA (Foreign Intelligence Surveillance Act) into law - that reauthorizes government, to spy on US Citizens without need for a warrant.

I interviewed Christine Bannon, US Public Policy Manager at Proton and she said this: “LLMs will be used by governments to sort through large data sets for intelligence, making it easier to conduct mass surveillance.”

We are honoured to welcome Saima Fancy, who has expertise in Data Privacy and its intersection with Cybersecurity and AI. We will discuss these new vulnerabilities introduced by LLMs and we’ll address new law that will now legally undermine individual’s data privacy and their civil liberties, and what are the implications of all this for emerging tech companies.

Saima Fancy has two decades of professional experience in chemical engineering, privacy engineering, law, data privacy and security. She is a speaker, mentor and active member and volunteer with the IAPP, All Tech is Human, Leading Cyber Ladies, Women in AI Ethics and Governing Council of U of T Faculty of Engineering.