What happens when attackers exploit AI’s rapidly growing capabilities, or when AI agents themselves find solutions that we neither intended nor can control?

Svenska Dagbladet interviews Aranya Consulting CEO Åsa Schwarz, KTH professor Pontus Johnson, and Martin Carlsson Wall about how AI is rapidly changing the cyber threat landscape and what consequences these developments may have for critical infrastructure.

Among other things, Åsa highlights the risk of attacks on critical infrastructure:

“If you want to cause major disruption, you could, for example, target train service, along with the subway. An extremely large number of people ride trains or the subway every day. If those services go down, it creates major disruption in all logistics. And that also causes widespread anxiety in society.”

She also points to a risk that is harder to assess—what happens if AI development becomes increasingly autonomous:

“The big problem is whether AI can develop AI on its own. If so, things could go really wrong, because then there would be no human to stop that development. Unfortunately, we’re not that far from that.”

This issue is becoming increasingly relevant as AI agents become more powerful and are given greater authority to act independently.

Svenska Dagbladet provides more details in the article ” AI Creates a New Threat.”