On December 8, European Union negotiators successfully reached an unprecedented agreement, marking the world's first comprehensive set of regulations for artificial intelligence (AI). The deal paves the way for legal oversight of technologies employed in widely used generative AI services, including ChatGPT, which promise to reshape daily life but have raised concerns about potential risks to humanity.
Negotiators from the European Parliament and the EU's 27 member nations overcame significant differences on contentious issues such as generative AI and facial recognition surveillance by law enforcement. This resulted in the signing of a preliminary political accord called the Artificial Intelligence Act. European Commissioner Thierry Breton celebrated the achievement, declaring the EU as the first continent to establish explicit guidelines for AI usage.
The breakthrough followed intensive closed-door discussions throughout the week, culminating in a marathon 22-hour session and continuing into a second round of negotiations on Friday morning. While the European Parliament is set to conduct a vote early next year, it is largely considered a formality with the deal sealed, according to Brando Benifei, the Italian lawmaker leading the negotiations. The law's full implementation is anticipated by 2025, accompanied by significant financial penalties for violations, reaching up to €35 million ($38 million) or 7 percent of a company's global turnover.
Generative AI platforms like OpenAI's ChatGPT have gained immense attention for their ability to mimic human creation in text, images, and music. However, concerns persist regarding potential effects on employment, privacy, copyright protection, and human safety. While other countries are also formulating their own AI regulations, the EU's robust guidelines could set a benchmark for global governance, according to legal expert Anu Bradford. Despite the milestone, concerns linger about rushed decision-making, the need for technical refinements, and potential loopholes in the finalized AI Act, particularly regarding systems used in migration and border control.