Tools of generative artificial intelligence pose risks to free and fair elections in a democratic society, stated European Commission Commissioner Vera Jourova.
According to her, the deepfakes misinformation poses a potential threat to the upcoming vote for the new European Parliament.
Jourová welcomed the safeguards informing users about the ‘synthetic origin of content’, but urged their further expansion.
“These efforts must be continued and intensified, given the high potential of realistic AI products to create and spread disinformation. The risks are especially high in the context of elections. Therefore I urge platforms to remain vigilant and to provide effective safety guarantees,” she warned.
The EU commissioner noted that she had met with OpenAI’s leadership and discussed these issues with them.
In April, a group of European lawmakers announced plans to develop a draft framework of rules to regulate the development of AI systems. Later, the European Parliament introduced amendments to the document that ban a broad range of uses of the technology.
These include restrictions on facial-recognition programs in public places and predictive policing algorithms. The bill also imposes transparency requirements on creators of generative models. They would be obliged to label content as AI-generated and to disclose information about the materials used for training.
In June, UN representatives said that AI-generated deepfakes undermine information integrity and fuel hatred in society.
Researchers at OpenAI warned of the growing threat of using language models to spread disinformation.
American scientists, for their part, urged countering deepfakes created by ‘hostile states’.