AI cyber attacks are a “critical threat”. This is how NATO meets them

Artificial intelligence (AI) plays a massive role in cyber attacks and is proving to be both a “double-edged sword” and a “major challenge,” according to NATO.

“Artificial intelligence allows defenders to scan networks and repel attacks more automatically, rather than doing it manually. But of course it’s the same game the other way around,” David van Weel, NATO’s deputy secretary general for emerging security challenges at NATO, told reporters earlier this month.

Cyberattacks, both on national infrastructure and on private companies, have grown exponentially and have become a focus of attention since the war in Ukraine. NATO said this year that a cyberattack on one of its member states could trigger Article 5, meaning an attack on one member would be seen as an attack on all and could trigger a collective response.

AI-based tools can be used to better detect and protect against threats, but on the other hand, cybercriminals can use the technology for more sporadic attacks that are harder to stop because there are so many at once.

AI can be used to try to break into networks using credentials and algorithms to crack systems, van Weel said.

Trying to solve the combinations “is a big challenge,” he said, adding, “We obviously want to be ethical users of AI.”

He said AI is being used for defence, “but of course we cannot guarantee that our adversaries attempting to break in will use AI in the same ethical way”.

“We have to take that into account in our defence. It’s definitely something we’re watching.”

Cyber ​​defense put to the test

How to defend against AI cyber attacks is being tested in Estonia’s capital Tallinn on the CR14 NATO Cyber ​​​​Range.

Earlier this month, army commanders from over 30 countries (not all NATO members) came to the Cyber ​​Range to test their skills on how they would defend their country while working alongside their allies.

Fictional storylines were created and one of the biggest challenges of the annual event was dealing with the threat of AI attacks.

“In the AI ​​experiment, it’s basically not a one-way street. It detects AI used by adversaries and, on the other hand, explores how AI can support our own operations,” said Bernd Hansen, Cyberspace Division Manager at NATO Allied Command Transformation.

“We open the technical experiments to the operating community to ensure that what we are trying to develop from a technical point of view actually serves the operator – ​​so that we do not march to the left when they want us to march to the right,” he told Euronews Next.

The exercises have helped the participants, but there is still a long way to go to counter the threat.

AI is “definitely a robust problem that I think is being addressed by the cyber community,” said Candace Sanchez, US chief planner, who attended the cyber exercises.

“But I think it’s going to take some time to really try to counter this threat. Working together in this effort to try to do this I think will help us move forward,” she told Euronews Next.

The Price of Internet Freedom

AI cyberattacks can be used not only to take down infrastructure but also to exploit information, said Alberto Domingo, technical director for cyberspace at NATO Allied Command Transformation.

“I think AI is a critical threat. The number of attacks keeps growing exponentially,” he told Euronews Next, adding that the world is just “living with these attacks” at the moment and needs more cybersecurity rules.

“We are not at a stage where we realize that this is simply not acceptable. These behaviors must not be allowed in cyberspace,” he said.

“It shows you that we still don’t have a collective common approach to respond to these things, but these things are just not acceptable.”

Though solutions are being worked on to combat AI cyberattacks, Domingo said we can’t stop them if we still want the internet to be a place of free thought and independence.

“We created cyberspace to be open to everyone. It is an environment for the development of ideas. We want that. We want freedom in cyberspace,” he said, adding that stopping it in favor of restricting what we can do online is too expensive.

“The price we have to pay for it [Internet freedom] is to be realistic and accept that there will be attacks on the networks,” he said.

“And the only way to deal with it is to use all mechanisms and all technologies to protect against these attacks, but also to react and recover from these attacks. I don’t think we can ever avoid them now.”