EU warns on AI: Loss of human control poses serious security risk

Henna Virkkunen, the European Commission's Executive Vice-President for Tech Sovereignty, Security and Democracy, has drawn attention to the security dimension of rapid developments in artificial intelligence technologies. Virkkunen stated that one of the most serious risks in AI is humans losing control over this technology.

Speaking to the Finnish public broadcaster Yle, Virkkunen emphasized that security measures for advanced AI systems must be kept constantly active. "These technologies are so advanced that it is extremely important that certain security measures are constantly in place and that control remains with humans," she said.

Virkkunen also recalled that some recent AI models have escaped the test environments created for them, causing harm in terms of cybersecurity. For this reason, she said that systems must be monitored not only before they are released to the market but also after they are put into use.

EU LAW AND EMPHASIS ON MONITORING

Stating that the EU is preparing through legislation against threats that may arise from AI, Virkkunen noted that the EU AI Act imposes an obligation on model providers to regularly assess risks and implement necessary security mechanisms.

Virkkunen expressed that there are actors in the world ready to use AI for malicious purposes, stating that this is a risk that must be constantly taken into account. "I wouldn't say I'm afraid of it, but of course it is a risk that is and must be constantly considered," said Virkkunen, noting that it must be ensured that AI models are safe.

Stating that global technology companies and other actors are taking the rules set by the EU into account, Virkkunen expressed that the most important tool against threats that may arise from AI is international cooperation. The European Commission Executive Vice-President emphasized that human oversight must be maintained at every stage of technological development.

News Source: 12punto