Interpol warns AI enhances cyber threats through speed and scale, not revolution

Artificial intelligence enables cyber threats by enhancing existing criminal tactics such as scams and fraud, according to Interpol.

AI is enhancing existing criminal techniques rather than revolutionizing them, Bjorn R. Watne, global chief information security officer at Interpol, told CNBC on the sidelines of Tech Week Singapore on Thursday. Scam operators can use AI to target multiple victims simultaneously, while improved translation and digital identities are making it increasingly difficult to distinguish fraudulent interactions from real ones.

"It's an evolution and not a revolution," Watne said, adding that AI is primarily increasing the "speed and scale" of existing techniques.

What companies should be protecting

As cyber threats evolve, companies should prioritize protective actions, according to Watne.

Instead, businesses should first identify their "crown jewels," the assets most critical to keeping their operations running, and determine who would be most interested in stealing or disrupting them.

They can then use threat intelligence to understand the tactics and technologies used by those adversaries and tailor their defenses accordingly.

"There is no need for everyone to try and protect against everything all at once when they're not being targeted by it," he said. The defenses needed can vary depending on whether a company faces opportunistic criminals or advanced persistent threat actors, he added.

Companies are also often hampered by a disconnect between senior management and their cybersecurity strategy, he said.

"There are still many industries where they haven't realized that everyone today is an IT company," Watne added.

While cyber risk is moving higher on corporate risk registers, cybersecurity has yet to fully make its way into the boardroom at many companies, he added.

Agentic AI raises new risks

Watne said he is particularly wary of agentic AI as the technology gains the ability to take actions on users' behalf, warning that mistakes could extend beyond incorrect information to real-world consequences.

"One thing is that an AI is feeding you the wrong information," Watne said. "But when an AI is actually performing an action and it starts doing the wrong actions, especially in the physical domain, that can have consequences on bodily harm of humans."

He pointed to the growing deployment of AI in devices such as cars and self-driving vehicles as an area of concern.

Watne said those risks are compounded by the relatively high level of trust people place in technology compared with areas such as financial services.

People tend to take a more control-focused approach to financial services, relying on safeguards such as PIN codes and remaining alert to risks such as card skimming, Watne said. When it comes to technology, however, they are often quicker to adopt new devices and apps without the same level of caution.

"When it comes to technology, the trust level is through the roof," Watne said, pointing to how readily people click through prompts and grant access to new technologies.

As AI systems become more capable of acting on users' behalf, he said that level of trust deserves greater scrutiny.