German authorities are warning that increasingly powerful artificial intelligence models could fuel a new wave of cyberattacks by making sophisticated operations easier, faster and more accessible to criminals.
Germany’s Federal Office for the Protection of the Constitution (BfV), the country’s domestic intelligence agency, and the Federal Office for Information Security (BSI) said Monday that they expect the cyberthreat landscape to deteriorate as AI capabilities continue to advance.
The BfV told German business daily Handelsblatt that large-scale cyberattacks are becoming easier to conduct as the barriers to entry fall. Artificial intelligence can assist attackers with reconnaissance, automate parts of cyber operations and help scale attacks more efficiently.
The BSI also warned that the misuse of AI is likely to increase. According to the agency, advanced models developed by leading technology companies demonstrated significant potential last year for identifying security vulnerabilities.
The capabilities have since progressed further.
The BSI said some AI models are now able to independently exploit vulnerabilities after identifying them, raising concerns that criminals could eventually use AI to automate increasingly complex attacks with limited human involvement.
Access to the most powerful AI systems remains constrained by their high computing requirements and costs. However, German officials expect that obstacle to become less significant as computing hardware becomes cheaper and more widely available.
That could potentially expand access to advanced AI capabilities beyond major technology companies and well-funded organizations, giving cybercriminals new tools to carry out attacks at greater scale.
German authorities stressed, however, that AI is not exclusively a threat to cybersecurity.
The BfV said artificial intelligence is also becoming an essential part of modern cyber defense, with the same technology that can help attackers potentially being used to detect threats, identify vulnerabilities and respond to attacks more quickly.
As AI models become more capable, authorities face a growing challenge: preventing criminals from exploiting those capabilities while ensuring that defenders can use the technology to stay ahead of increasingly sophisticated cyber threats.