AI Regulation Gap: 15% of Jurisdictions Ready for 2026

It’s wild to think that 70% of financial institutions are actively using or exploring AI, but almost no global financial regulations actually address it. This gap is creating real problems for financial stability and consumer protection. How are we supposed to encourage innovation when the rulebook for these intelligent algorithms and predictive models hasn’t even been written yet?

Only 15% of Jurisdictions Have Specific AI-Focused Financial Regulations

A 2025 survey from the Financial Stability Board (FSB) paints a pretty stark picture. We’re looking at a world where only about 15% of jurisdictions have bothered to enact specific, dedicated AI regulations for their financial sectors. That figure is way too low, especially when you see how fast AI is being adopted in banking, asset management, and insurance. Most regulators are just stretching old, analog-era rules to cover new digital tools or they’re issuing broad ethical manifestos with no real teeth. This patchwork leaves huge gaps in critical areas like algorithmic bias, model explainability, and the systemic risk that comes from interconnected AI systems. I’ve seen firsthand how banks get completely lost trying to apply outdated rules to a new machine learning model for credit scoring. The ambiguity leads to inconsistent application of consumer protection laws and creates openings for firms willing to cut corners.

The EU’s AI Act Sets a Precedent for Risk-Based Governance

While most of the world is lagging, the European Union’s AI Act, which was formally adopted in late 2024, is a major step forward. Though the legislation is broad, it does something very smart: it categorizes AI systems by risk level. Many financial service applications fall into the “high-risk” bucket, which means they’re subject to tough requirements like mandatory human oversight, strong data governance, and transparency. A European Commission press release confirms the goal is to make sure AI is safe and respects fundamental rights. In my opinion, this risk-based approach is the only pragmatic way to go. It accepts that some AI applications are more dangerous than others and focuses regulatory firepower where it’s needed most. The big challenge now is for financial regulators to issue detailed guidance on how to apply these high-level principles to things like complex derivatives pricing or automated portfolio management without killing innovation in the process.

Cross-Border Data Flows Complicate Harmonization Efforts

Finance is global, which means AI models get trained and used across different countries. That’s a problem when every jurisdiction has its own data privacy laws, whether it’s the EU’s GDPR, California’s CCPA, or Brazil’s LGPD. A Reuters report on data regulations has already pointed out how complex this is getting for multinationals. This patchwork of data protection rules is a massive roadblock for harmonizing AI regulation. Just think about it: an AI system developed in New York using local customer data, then deployed by a subsidiary in Frankfurt. The rules on data governance, consent, and even what counts as “personal data” can be completely different. This directly impacts the ability to build globally consistent AI models and share what works. Regulators have to get out of their national silos and start working on international data transfer agreements and common data ethics. Without that, any talk of standardizing AI governance is just talk.

Basel Committee is Developing AI Risk Management Principles

With all this chaos, at least the international bodies are starting to wake up. The Basel Committee on Banking Supervision (BCBS), which is the main standard-setter for banking, has been working on what it expects from banks using AI and machine learning. You can see their thinking in discussion papers from the Bank for International Settlements (BIS), which zero in on principles for solid risk management covering model validation, data quality, governance, and ethics. This is good news, because what the BCBS says usually ends up in national banking regulations around the world. These will be principles, not strict prescriptions, but they’ll give national supervisors a foundation to work from. My concern is speed. Can these principles be adopted consistently and quickly enough? The deliberative process of these international bodies is always playing catch-up with the tech.

A Lack of Unified Global AI Ethical Guidelines

The deepest problem for regulatory harmony might be the simple fact that there’s no universal agreement on AI ethics. Lots of countries and organizations have published their own principles, but they all reflect different cultural values and legal traditions on concepts like fairness, accountability, and human autonomy. For instance, what’s considered “fair” in an algorithm that decides who gets a loan can be viewed very differently from one society to the next. A Pew Research Center report showed that even the experts can’t agree on how to build ethical AI. This divergence makes it almost impossible to create harmonized regulations. If we can’t agree on the basic ethical guardrails for AI, how can we possibly agree on the technical standards to implement and oversee it? This directly impacts how AI models are designed and trained, influencing everything from data selection to how they’re audited. Without a common ethical compass, real global alignment is a fantasy.

Challenging the Conventional Wisdom: More Regulation Is Not Always the Answer

The knee-jerk reaction to the complexity of AI in finance is often to demand more regulation. The conventional wisdom is that a complete, prescriptive global rulebook is the only way to manage the risks. I think that’s a mistake. Of course, we need targeted, risk-based rules. But an overly prescriptive approach will just choke innovation and create crushing compliance costs that hit smaller fintechs the hardest. AI technology moves so fast that any super-detailed regulation will be obsolete before the ink is dry. I think we need to focus on meta-regulation instead: set clear principles for governance, accountability, and transparency, and back them up with strong supervision and industry-led best practices. Regulators should be mandating strong internal controls, requiring tough model validation, and demanding explainability where it counts, not trying to write the rules for specific algorithms. The goal should be a culture of responsible AI development, where firms can innovate within defined risk boundaries, not a world where they’re trapped by an exhaustive list of prohibitions that won’t apply to tomorrow’s tech. It’s a much more realistic path forward.

It’s a complicated path toward getting AI regulation right in finance, full of friction between countries and technology that won’t slow down. But we have to find a way to encourage smart innovation without putting the global financial system itself at risk.