Artificial intelligence technologies have boomed over the last few years. But despite all that growth, the industry faces little regulation. In the European Union, that could soon change. On Wednesday, EU lawmakers signed off on the world’s first set of comprehensive rules for artificial intelligence: the Artificial Intelligence Act.
AI opens new possibilities—sometimes in unexpected ways. This week, musician Paul McCartney announced that AI would allow the Beatles to release a new song. The classic rock band broke up in 1970, and two of its members have died. Now modern technology enables sound engineers to clean up and re-mix old demo recordings.
But not every new possibility is so innocent. Governments can use AI-powered facial recognition to track citizens and catch criminals—sometimes arresting the wrong person. . AI products targeted toward children could expose kids to harmful content. The tech has grown so fast that it’s hard to know all the risks. That’s why some experts recommend a pause on AI development.
The EU’s Artificial Intelligence Act aims to strike a difficult balance. Lawmakers want to prevent the harms of AI. But they still must leave room for companies to innovate.
Boniface de Champiris works for the Computer and Communications Industry Association, a lobbying group for tech companies. “Europe’s new AI rules need to effectively address clearly defined risks, while leaving enough flexibility for developers to deliver useful AI applications to the benefit of all Europeans,” he says.
If passed, the new laws would govern any product or service that uses AI. The act would classify AI systems by four levels of risk, from minimal to unacceptable.
Minimal-risk systems include products such as video games and email spam filters. Higher risk systems include those used for employment and education. These technologies can have a huge effect on peoples’ lives.
The act would ban some uses of AI entirely. People couldn’t wield AI for “social scoring” systems to judge citizens based on behavior. Governments within the EU couldn’t use AI for “predictive policing.” Predictive policing tools identify potential lawbreakers before they’ve even broken a law. Imagine getting arrested for a crime a computer says you might commit!
The new rules could have an impact beyond the EU. EU regulations tend to become global standards. Countries such as the United States might someday adopt similar laws.
When will these laws take effect? It could take years. The EU’s member countries, the Parliament, and the European Commission must negotiate before the act passes. Those negotiations could bring further changes. Experts expect final approval before the year’s end. But then comes a “grace period,” giving companies time to adapt. That could take two more years.
New technologies and new laws have something in common. Both can bring unintended consequences. If lawmakers get this wrong, AI companies might not want to work in the EU. But if they get it right, companies will feel the freedom to innovate—and consumers will feel safer.
Now therefore, O kings, be wise; be warned, O rulers of the Earth. — Psalm 2:10
