By the end of April, the European Parliament will set its sights on a list of prohibited practices, including social scoring, predictive policing, algorithms that scrape photos indiscriminately from the internet, and real-time biometric authentication in public spaces. I was. But on Thursday, conservative European People’s Party MPs still questioned whether the biometric ban should be withdrawn. Some political forces and groups see it as a force to fight crime, while others, like the Progressives, see it as a system of social control,” says co-rapporteur Brando Benifei, an Italian. His MEP in the socialist and democratic political groups.
Next, there was talk about the types of AI that should be flagged as high risk, such as algorithms used to manage employees in companies and algorithms used by governments to manage transitions. They are not prohibited. “But I underline the words because of their potential meaning. potential— Regarding our rights and interests, they have to go through some compliance requirements and make sure those risks are properly mitigated. Developers must show the data they used to train the AI and demonstrate how they actively tried to eliminate bias. A new AI agency will also be set up to create a central hub for enforcement.
Companies deploying generative AI tools such as ChatGPT are required to disclose whether their models are trained on copyrighted material, increasing the likelihood of litigation. Text or image generators, such as MidJourney, must also identify themselves as machines and mark their content in a way that indicates that it is artificially generated. Additionally, tools must not generate child abuse, terrorism, hate speech, or any other type of content that violates EU law.
One person, who requested anonymity to avoid attracting negative attention from lobbying groups, said some of the rules for general-purpose AI systems were watered down in early May following lobbying by tech giants. Stated. The requirements of the underlying model underlying tools like ChatGPT are now audited by independent experts.
But Parliament has agreed to register foundation models in a database before they are put on the market, so companies will have to tell the EU what they have started selling. “This is a good start,” says Nicolas Moës, director of European AI governance at think tank Future Society.
Lobbying by big tech companies, including Alphabet and Microsoft, is something legislators around the world need to be wary of, says Sarah Myers West, managing director of the AI Now Institute, another think tank. “I think we are starting to see new strategies for how they are trying to tilt the policy environment in their favor,” she says.
What the European Parliament ended up with is a deal that will please everyone. “This is a real compromise,” said one member of parliament, who spoke on condition of anonymity because he was not authorized to speak publicly. “Everyone is equally unhappy.”
