How AI is expanding the cyber power of nation states

AI has yet to produce the entirely new form of catastrophic cyber attack once forecast by researchers. Its immediate impact is already visible elsewhere: stages of an attack that once took days can now be completed within minutes, while language barriers that exposed phishing campaigns are disappearing.

Dr Gil Baram is a cyber strategy and policy expert with more than 15 years of experience. She is a senior lecturer at Bar-Ilan University and a non-resident research scholar at the Center for Long-Term Cybersecurity at the University of California, Berkeley. A former Fulbright cybersecurity postdoctoral fellow at Stanford University, she has also held cyber intelligence roles within Israel’s National Information Security Authority.

In this exclusive interview with the Cyber Security Speakers Agency, Gil examines how AI is increasing the speed and reach of cyber attacks, how nation-state actors are using large language models and where accountability should sit when AI enters cyber systems. She also explains why offensive cyber capabilities now function as instruments of geopolitical power and deterrence.

Q. What does the current evidence reveal about AI’s impact on the methods, speed and scale of cyber attacks?

Gil Baram: “To understand the way AI changes the scale and speed, we need to go back and remember that, a year or two ago, many researchers thought or estimated that when AI was used, we would see new types of cyber attacks. They anticipated something completely new that we had not seen before, perhaps a catastrophic attack involving a new method or new techniques.

“The interesting thing is that we have not seen something like that yet. We do, however, see a change in both speed and scale. The speed of attacks is getting much faster because many parts of the attack kill chain are being automated by AI. Some parts that used to take several days or hours of human work are now taking minutes. That changes a lot.

“We also see a big change in the scale of attacks. Even nation-state attackers, such as North Korean attackers, are getting better in two ways. One is that the language barrier basically does not exist any more. Before the use of AI, North Koreans had to use tools such as Google Translate. 

“The translation was not as good, and it was easier to identify phishing emails or scams. They now have perfect English, which has enabled a larger scale.

“They have also got better in their use of technology. We have seen reports from Google and OpenAI describing how nation-state actors are using large language models to learn new techniques, including new coding and attack techniques. They are getting better, and the scale is larger than before.”

Q. Where should accountability sit when AI is deployed in cyber systems, and what level of transparency is required?

Gil Baram: “The two most important things, the way I see it, are accountability and transparency.

“We have to remember that human beings, for now, are still accountable for what is going on, what the results are and how the technology is used. That comes hand in hand with transparency. We have to be transparent about what goes inside the machine, what comes out and how we use it.

“We can use AI for many things, but we have to think: do we want to do that? Do we want to take the human perspective out and put artificial intelligence inside? I think these are the main principles.”

Q. How have offensive cyber capabilities become instruments of geopolitical power and deterrence?

Gil Baram: “Cyber security has been geopolitical for a long time. We tend to look at cyber and AI as mainly technical subjects. We also tend to think that, if we are not computer scientists, we are not the ones who should understand what is going on.

“The truth is that offensive cyber capabilities have been used for years by states, nation states and non-state actors. They are tools of power and deterrence, extending far beyond their technical functions.”

Q. What do you want audiences without a technical background to understand about cyber and AI threats?

Gil Baram: “The main thing I hope audiences take away, and that is my passion and goal, is to understand that cyber and AI threats are relevant to our everyday lives.

“They are not something out there that we only read about in the news. They are relevant to everyday life, and we do not have to be technical people to understand them or protect ourselves. That would be my main message.”

This exclusive interview with Gil Baram was conducted by Tabish Ali of the Motivational Speakers Agency for European Business Magazine.