The following is a guest editorial courtesy of Eric Odotei, Group Head of Regulatory Reporting at Finalto.

AI is no longer knocking at the door. It has walked in, taken a seat at the table, and started rewriting the rules. It is no longer the technology of the future; it is here now. It powers our social media feeds, search engines, influences investment decisions, drives personalised recommendations, and even helps doctors diagnose illnesses faster. The shockwave of its impact is rippling through every industry, and financial services are no exception. Indeed, in a heavily regulated sector like financial services, the emergence of AI-powered systems promises to be especially disruptive.

In my previous article, I explored how financial services firms can more effectively incorporate AI into their operations, and the importance of an AI strategy that contributes to transparency, fairness and compliance.

This month, I want to take a step back and consider the broader regulatory landscape. Few would argue against the need for some form of AI oversight. But the real challenge lies in finding the balance between effectively managing risk without stifling innovation. AI is growing fast, but in many ways, it is still a toddler. It’s learning, adapting, and evolving in real time. The truth is, we don’t fully understand what it’s capable of, or indeed what it might become.

Minefield vs level playing field

On one hand, there is a clear need for some level of oversight. AI systems are not neutral. They learn from data, and if that data contains bias, the AI will carry those biases forward, often at scale. AI can also be exploited to spread misinformation, and with the rise of deepfakes, it poses a serious threat to truth and authenticity. AI is capable of making decisions that affect our lives without explaining how those decisions were reached. In finance, healthcare, and transport, that lack of transparency poses real risks. Regulators across the world are working to introduce rules to make sure AI is used responsibly.

But there is another side to this story. In order to tame the beast, one must understand its nature and purpose. This takes time and if we rush in with too many rules, we risk curtailing the freedom of innovation before it reaches its full potential. Heavy regulation and compliance requirements can make it harder for start-ups and smaller firms to compete, leaving the field dominated by big technology companies with the resources to manage the costs. Businesses may even hold back from adopting AI altogether, worried about legal implications and ultimate consequences. Jurisdictions that take an overly strict approach could also find themselves falling behind those that are more flexible.

Risk vs opportunity?

This is not a hypothetical concern. The European Union’s General Data Protection Regulation was a big step forward for privacy, but critics argue it also created barriers that slowed digital innovation compared to other regions. Now, the EU has introduced the AI Act, the world’s first comprehensive law designed to govern artificial intelligence. It adopts a risk-based framework, classifying AI systems into categories such as unacceptable, high, limited, and minimal risk. High-risk applications, such as those in healthcare, education, and law enforcement, will be subject to strict requirements, including transparency, human oversight, and conformity assessments. While the aim is to protect citizens and ensure ethical AI, some fear that the cost of compliance and the complexity of implementation might discourage smaller firms from innovating and could inhibit Europe’s ability to compete globally. Many are asking whether history is about to repeat itself.

Different regions are taking slightly different approaches. While the EU is moving ahead with a risk-based model that puts strict rules on high-risk AI applications. The United States is taking a lighter approach, relying on voluntary codes, sector-specific guidance, and an innovation-first attitude. The United Kingdom has introduced regulatory sandboxes and invested in AI safety testing, giving businesses room to experiment without heavy restrictions. China has gone for strong central control, with strict rules and requirements for alignment with state policies. For instance, China’s Ministry of Public Security announced last month that it will collaborate with the Ministry of Industry and other regulatory authorities to strengthen oversight and introduce tighter controls on assisted driving systems. Countries like Canada, Australia, and India are opting for more adaptive frameworks that mix flexibility with regulation.

Finding the balance

Which approach is right? The truth is, there is no simple answer. What is clear is that neither extreme works. Too little oversight creates chaos, and too much can choke progress. The best way forward is a balanced approach, built on adaptive, risk-based rules that can evolve as the technology evolves. That means focusing on higher-risk sectors like finance and healthcare, demanding transparency and explainability for AI models that affect critical decisions, and using tools like regulatory sandboxes to test new ideas safely rather than banning them outright. It also means regulators and industry working together on co-regulation and shared accountability.

The debate around AI regulation is not only technical or legal, it is about values. How do we balance progress with responsibility? Do we prioritise speed over safety, efficiency over ethics, profit over people? In the end, regulation is about trust. Without trust, people will hesitate to adopt AI, and innovation will slow. With trust, AI has the potential to deliver real benefits for businesses and society.

The paradox is not going away. The challenge is not to choose between two extremes, but to find a workable middle ground, one that encourages innovation while safeguarding the people AI is meant to serve. How we strike that balance will shape not only the future of AI, but also the future of our economies and societies.

For now, each jurisdiction is still coming to terms with the nature and potential of AI and we need the space to remain flexible, to experiment responsibly, and above all, to collaborate across borders, sectors, and disciplines. The question is no longer whether AI should be regulated, but how we can regulate it wisely, without undermining the very innovation we are seeking to harness.