From The Wall Street Journal:
European lawmakers reached a political deal on regulating artificial intelligence, marking a step toward establishing a comprehensive AI law in Western countries.
The European Union’s Artificial Intelligence Act would be the latest in a series of regulations the bloc has pushed forward that is expected to reverberate through the global tech industry and affect some of its biggest players. Earlier legislation from the EU set out new competition and online content rules that affect large U.S. tech companies including Meta Platforms, Apple and Google’s parent, Alphabet.
“The EU is the first in the world to set in place robust regulation on AI,” said Dragos Tudorache, a member of the European Parliament from Romania who was one of the body’s lead negotiators of the AI Act.
The deal agreed to by lawmakers includes bans on several AI applications, such as untargeted scraping of images to create facial-recognition databases, and sets rules for systems that lawmakers consider to be high-risk, according to a statement from the European Parliament. It also includes transparency rules for general-purpose AI systems and the models that power them.
Penalties for breaking the rules could reach up to 7% of a company’s global revenue, depending on the size of the company and the rule that was broken, according to the statement.
Europe’s AI Act was first proposed in 2021, before chatbots such as OpenAI’s ChatGPT and Google’s Bard were publicly launched. The introduction of those chatbots and the popularization of a range of other AI applications in recent years put new scrutiny on the legislation and prompted some lawmakers to seek to introduce new provisions.
One of the most controversial aspects of the European legislation was determining whether to set blanket rules for general-purpose AI and so-called foundation models, which are trained on vast data sets and underpin more-specialized AI applications.
The AI Act will require those systems to follow a set of transparency rules that include complying with EU copyright law and creating detailed summaries about the content used to train AI models, according to the statement from the European Parliament. High-impact models that are deemed to create systemic risk will face tougher rules, including a requirement to assess and mitigate such as risk.
. . . .
The deal faced swift criticism from industry and consumer groups. DigitalEurope, a tech lobby group, said the new rules would be expensive for AI companies to comply with and risk putting Europe at a disadvantage. “The AI race is not one Europe can miss out on,” said Cecilia Bonefeld-Dahl, the group’s director-general.
. . . .
At The New York Times, Adam Satariano writes, “even as the law was hailed as a regulatory breakthrough, questions remained about how effective it would be.
“Many aspects of the policy were not expected to take effect for 12 to 24 months, a considerable length of time for AI development. And up until the last minute of negotiations, policymakers and countries were fighting over its language and how to balance the fostering of innovation with the need to safeguard against possible harm.”
So while an agreement has been reached, its precise outlines—and how firm they may be—remain to be examined and understood.
Link to the rest at The Wall Street Journal (Sorry if you encounter a paywall)
PG says this is a mistake that will kill a great deal of AI research and development in the EU. He predicts that many bright AI researchers and developers will decide to head to employment opportunities outside of Europe.