European Union officials reached a deal on Friday regarding a comprehensive bill aimed at regulating artificial intelligence in their first attempt to broadly regulate this technology. The agreement was reached after three days of negotiations in Brussels, including a 22-hour session from Wednesday to Thursday, and a copy of the agreed-upon law has not yet been obtained.
Key Points
EU lawmakers reached an agreement on Friday regarding the key elements of the AI Act, which will be the first major law to regulate artificial intelligence technology in an attempt to allow its beneficial use while ensuring protection from risks.
Details
Previous drafts of the law aimed to restrict how artificial intelligence is used by law enforcement, government, and service companies such as water and energy providers, and stipulate transparency requirements for major AI companies like OpenAI, Microsoft, and Google. It will also require labeling images created by artificial intelligence to indicate so, which is something the United States is also working to implement.
Impact
The agreement makes the EU the first governmental body to pass comprehensive regulation on artificial intelligence. However, the agreement still requires final approval as it must be voted on by the 27 member states, and the technical details need to be completed, according to the New York Times report, so it is unclear when the vote will take place.
Important Quote
“Historic!” European Commissioner Thierry Breton wrote on Twitter on Friday. “The AI Act is more than just a rulebook – it is a launchpad for startups and researchers in the EU to lead the global AI race. The best is yet to come!”
What to Watch For
What has been included in the agreed bill. Reuters reported that major countries such as Germany and France were challenging in discussions on Wednesday, including proposals that involve requirements for training system documentation and demonstrating that they have taken steps to mitigate risks and conduct audits, saying they prefer to allow companies to regulate themselves.
What We Don’t Know
How the AI Act will be enforced. It is expected that legislators from 27 countries will be involved and it will require new hires, according to the New York Times report, and it is likely to face legal challenges from major tech companies before it goes into effect.
Background
The EU began working in 2018 on what would become the AI Act, making it a leading regulatory body for this emerging technology. The first version was available three years after work began, according to the New York Times report, but it did not address “general-purpose” models like those behind ChatGPT and Google’s Bard, due to the rapid pace of technology development. The United States has also worked on regulating artificial intelligence, but Congress has not yet united behind any comprehensive measures like the AI Act. The White House issued an executive order in late October outlining “the most comprehensive actions ever taken to protect Americans from the potential risks of artificial intelligence,” which requires some companies to share safety test results and other information with the government, among other things.
Leave a Reply