INTERPOL is warning that artificial intelligence is fueling cyber threats because it strengthens existing criminal methods such as scams.

In an interview on Thursday during Singapore Tech Week, INTERPOL Global Chief Information Security Officer Bjorn R. Watne said AI is an enhancement of existing criminal methods rather than a complete transformation. Fraudsters can use AI to target multiple victims at once; at the same time, improvements in translation technology and digital identity verification are making it increasingly difficult for people to distinguish fraudulent interactions from genuine ones. "This is an evolution, not a revolution," Watne said, adding that AI is mainly boosting the "speed and scale" of existing techniques.

What companies should prioritize protecting

Watne noted that as cyber threats continue to evolve, companies should prioritize protective measures. Companies should first identify their "crown jewels"—the assets that are critical to maintaining operations—and determine which actors are most likely to try to steal or destroy them. After that, companies can use threat intelligence to understand the tactics and techniques used by these adversaries and tailor their defenses accordingly. "If you are not the target of a specific threat, you do not need to try to defend against every type of threat at once," Watne said, adding that the defenses required differ depending on whether a company faces opportunistic criminals or advanced persistent threat (APT) attackers. He also mentioned that a disconnect between senior management and cybersecurity strategy often hinders companies' efforts. "There are still many industries that have not realized that every company today is actually an IT company," Watne added. He also pointed out that although cyber risk is rising in importance on corporate risk lists, at many companies cybersecurity issues have not yet truly entered the boardroom's field of view.

Agentic AI brings new risks

Watne said that as such AI technologies gain the ability to take actions on behalf of users, he is highly cautious about this technology. He warned that when things go wrong, the consequences may go beyond providing incorrect information and could even lead to serious real-world consequences. "It is one thing for AI to give you incorrect information," Watne said, "but when AI actually executes actions and takes the wrong action—especially in the physical world—it can cause physical harm to people." He pointed out that the widespread application of AI in devices such as cars and autonomous vehicles is one area of concern. Watne noted that people's trust in technology is relatively high compared with areas such as financial services, which further amplifies the related risks. He said that in financial services, people tend to take a more control-oriented approach, relying on safeguards such as PIN codes while remaining vigilant about risks such as card theft; however, when facing new technologies, people are often more willing to adopt new devices and apps quickly without the same caution. "When it comes to technology, people's trust is astonishingly high," Watne said, noting that people often click on prompts without hesitation and grant access permissions to new technologies. He said that as AI systems become increasingly capable of taking actions on behalf of users, this level of trust deserves deeper scrutiny.