The EU’s Artificial Intelligence Act is causing ripples across the globe, reshaping how companies from Washington to Tokyo handle AI governance. New research indicates that Brussels is setting a standard that transcends its own borders, affecting any organization whose AI systems touch the EU or impact its citizens, businesses, and public institutions. This Act has been officially in effect since 2024, but its requirements are being rolled out gradually. By December 2025, bans on the riskiest uses of AI and transparency rules for general-purpose AI models were already in place. However, it’s the regulations for high-risk systems—including hiring, credit, and healthcare—that will become fully binding by August 2026.
Now, let’s break it down a bit. The Act carries significant weight, with penalties for serious breaches hitting €35 million or 7% of a company’s global annual revenue. But don’t get it twisted—while there’s some buzz about companies recognizing AI use, the reality is stark. A mere 13% of companies across all sectors have any formal AI governance framework, regardless of whether they mention the EU AI Act. Of that small fraction, just over half, at 53%, specifically reference the Act. And get this—47% of those companies are actually based outside of the EU.
The level of engagement with the Act varies widely depending on the industry and geography. Information technology firms are leading the charge, accounting for nearly 40% of all non-EU companies that cite the Act. Communication and financial services add another 29% to that mix. Regionally, North America stands out, with just under 40% of non-EU engagement driven largely by U.S. tech and healthcare firms that have a notable presence in the EU market. Following closely are companies from the UK, Switzerland, and Norway—about 24%—who maintain strong commercial ties with the bloc. Meanwhile, Asian firms are also getting in on the action, citing the Act at a rate of roughly 28%, particularly those entrenched in global AI supply chains.
Now, here’s where it gets interesting—the U.S. showcases a striking contrast with its generally hands-off approach to AI regulation. While there’s no overarching federal AI law in the States, American firms are still responsible for 35% of all non-EU citations of the EU Act, making them the largest national contributor. Within the U.S., 53% of the companies citing the Act come from the information technology sector, and one in five U.S. IT companies in the dataset references it, the highest rate of any sector across the nation. Major tech players like Microsoft, Google, OpenAI, and xAI have even voluntarily aligned with pieces of the EU’s AI Code of Practice, largely motivated by the desire to maintain access to European markets.
Those companies that do mention the Act generally have their bases covered. They boast clear AI plans, board-level oversight, and transparency regarding the data they use. Interestingly, non-EU firms referencing the Act actually outshine EU companies in this regard. However, when it comes to workforce training, EU firms lead the way, with nearly 50% offering reskilling or AI literacy programs compared to about 40% of their non-EU counterparts.
But let’s not get too comfortable. Strategy oversight is one thing, but monitoring what AI systems are actually doing on a case-by-case basis is another beast altogether. A staggering 12.4% of companies worldwide have a policy requiring human review of individual AI decisions, and nearly half of them haven’t figured out how to implement that in practice. Rights checks? Even rarer—fewer than one in four companies assess whether their AI systems could infringe on employee rights, even among those most engaged with the Act. But hold on, that’s about to change. Starting August 2026, firms employing high-risk AI in hiring, credit, or healthcare will be legally obligated to conduct those checks prior to rollout, and they’ll need to report their findings to regulators.
So, what’s next for companies navigating these new waters? Will they adapt swiftly enough to meet the looming deadlines, or will the penalties prove to be a wake-up call? Only time will tell…
Kaynak: Orijinal Haber