This article was first published on LinkedIn – follow us there for regular updates on all things tech recruitment.
If you’re in banking and wrestling with how to roll out AI, chances are you’ve felt the tension: how much oversight is too much? That question is getting very different answers depending on whether you’re operating under U.S. or European regulations.
Let’s break it down - because on one side of the Atlantic, there’s a rulebook that’s already in print. On the other, the pages are still being written, with a whole lot of red pens in hand.
Europe’s AI Act: Strict, Structured, and Here to Stay
The EU isn’t messing around. With the AI Act now in effect (since August 2024), Europe has drawn a bright line in the sand. AI systems are classified into risk tiers—“unacceptable,” “high,” “limited,” and “minimal.” And if your AI falls into the “high-risk” bucket (think credit scoring, fraud prevention, or anything involving sensitive data), expect heavy scrutiny.
Banks there must prove their algorithms are safe, unbiased, and explainable. That means transparency reports, human oversight, quality data, and the ability to explain how your model got to its decisions. And yes, non-compliance could cost you up to €35 million or 7% of global revenue—whichever hits harder.
So while it’s rigid, you know where you stand. That clarity? It’s what EU regulators are betting will build long-term trust in AI, especially for sensitive sectors like finance.
The U.S. Approach: Looser, Livelier, and (Maybe) Riskier
Now, swing over to the U.S., and things look very different. There’s no single national AI law. Instead, it’s a mosaic: the CFPB is focused on fairness in lending, the FTC keeps an eye on deceptive practices, and various states are pitching their own rules. Add to that the recent shift in federal tone, with the Biden-era AI executive orders now rolled back under the Trump administration, and you’ve got a regulatory wild west.
The industry voice here? Mostly, “Keep it light.” Folks like Senator Ted Cruz are pushing for AI innovation sandboxes where banks can experiment without heavy-handed rules. The ABA (American Bankers Association) is also chiming in, saying: “Let’s focus on the use of AI, not just the tech itself. Banks already know how to manage risk—give us room to breathe.”
That breathing room, though, can come with consequences. Without a unified framework, U.S. banks may struggle with consistency. One state might flag your AI model, another might give it a pass. That uncertainty? Not great when you’re trying to scale responsibly.
What Does It All Mean for Banks?
Here’s where it gets interesting:
- Compliance Juggling Act: If you’re running operations on both continents, you’re essentially playing two different regulatory games. One with clear boundaries (EU), and one with a moving target (U.S.).
- Innovation vs. Caution: Europe’s strict guardrails may slow things down, but keep risks in check. In the U.S., you can move faster, but the lack of rules could lead to missteps.
- Who Sets the Standard?: There’s a bigger question brewing here—who becomes the global leader in ethical AI for finance? The EU’s bet is on structured trust. The U.S.? Speed and market power.
So, What Should You Do?
If you’re in banking leadership, now’s the time to stay sharp. Monitor these frameworks, invest in flexible AI governance, and build explainability into everything. Whether you’re in New York or Frankfurt, it’s not about choosing sides—it’s about being ready for both.
The smart money says we’ll eventually see some convergence between the two approaches. But until then, your compliance team and data scientists better be on good terms.
This is where Trinnovo Group comes in -
No matter what stage of the AI/ML transformation process you’re at, our Financial Services team can provide the end-to-end solution be it advisory or augmentation.
Contact us directly to learn more about our market-leading service:
Me: Senior Principal Data Consultant at Trust in SODA
Francis.alexander@trustinsoda.com
Connor Nurse: Head of US, Risk, Compliance & Finance Specialist at Broadgate
James Davis: Principal AI/ML Consultant at DeepRec.ai