Why we need a pro-innovation approach to AI in financial services and beyond

Artificial Intelligence (AI) is the world’s next transformative technology, offering unprecedented opportunities for productivity and growth. However, instead of embracing the transformative power of AI, some governments are reacting with excessive caution. This is often driven by fear and zero-sum thinking that fails to grasp the massive potential of AI to create entirely new markets and services. A pro-innovation approach, especially in financial services, is essential if we are to unlock AI's full potential and avoid stifling the growth it promises.

AI technologies are already transforming financial services by improving cost-efficiency and better managing risks. Companies like Klarna have started integrating AI into their processes, allowing them to halve their workforce while maintaining efficiency. This is only the beginning. AI applications in the financial sector—from fraud detection to marketing—are setting the stage for disrupting operating models. 

Yet, the true breakthrough will come not just from improving efficiency but from radical innovations that AI can make possible. This revolution requires investment beyond automation of existing processes in existing firms. We must encourage development of brand new AI applications and new types of firms. New, unknown applications, opening untapped markets that will fundamentally alter industries including financial services. However, for this to happen, we need to support, not constrain, the development and deployment of AI.

The most exciting applications of AI are yet to come

The true power of AI is not just in making existing systems more efficient but in creating radical innovations that we can’t yet fully imagine. While sustaining and disruptive innovations already drive growth, AI holds the potential to break new ground by enabling services that have never existed before. These "unknown unknowns"—markets we can’t yet anticipate—are where AI will truly shine. 

For instance, think of big service gaps in the market where it hasn’t to date been economic for humans to serve everyone: like private tuition, private banking, personal assistants, or personal trainers; all made affordable and scalable by AI. We are on the brink of a future where AI enables everyone to access services previously reserved for a few. To get there, we need investment that focuses on discovering new applications of AI, not just refining existing technology.

Tax incentives and research funding should include the discovery of new business models, not just improving foundational models. The financial and societal returns from these new applications could be enormous, particularly in sectors where the human element is currently a bottleneck.

Fear is driving a rush to regulate, based on zero-sum thinking

Unfortunately, fear of AI’s potential—whether based on social or existential concerns—has triggered calls for over-regulation, especially in Europe. This reaction echoes the "rage against the machine" that accompanied earlier technological revolutions. The same zero-sum thinking that assumes machines replace humans without creating new opportunities is behind many of the calls to limit AI's development.

In reality, AI can drive productivity growth by creating more with less, lowering prices for goods and services, and freeing up spending power to fuel new industries. This has been the pattern with every major technological advance, and there is no reason to believe AI will be any different. The zero-sum fallacy—the idea that there is only so much work to go around—simply doesn't hold. AI will create new jobs, new industries, and a larger, more prosperous economy. But this growth is contingent on allowing AI to flourish, rather than being hemmed in by overzealous regulation.

Similarly, the fear of AI posing a catastrophic risk often stems from speculative scenarios that overestimate the autonomy and capabilities of current models. AI science still has a long way to go and researchers are rightly exploring alternative architectures and more efficient frontier models. The rush to regulate AI, however, results in definition problems and overreach. Most technologies considered AI a few years ago would not be considered so today and, if trends persist, it is doubtful we will consider today’s models real AI in a few years.

The European Union's AI Act, which equates risk with the scale of model computation, is an example of regulatory overreach that could stifle innovation. Risk should be evaluated based on applications, not the size of the AI model. A blanket regulatory approach could prevent new technologies from emerging and impede the economic gains AI can offer.

We need a pro-innovation approach to AI

AI, like all revolutionary technologies, needs smart governance, not stifling control. It is crucial to strike a balance between encouraging innovation and addressing legitimate safety concerns. However, regulations must be sector-specific, not blanket measures that apply across the board. The financial services sector, for example, already has advanced frameworks for managing risks associated with algorithmic decision-making, and these can be adapted to account for AI's growing role.

The UK's Department of Science, Innovation, and Technology (DIST) is taking a more pragmatic approach, focusing on tweaking existing sectoral regulations rather than reinventing the wheel. This approach allows for responsible AI adoption without freezing innovation. Other countries should follow suit. 

When it comes to existential risks—such as the potential misuse of AI by bad actors—these are best addressed through specialist agencies like intelligence and defence departments, not through generalist regulations that hinder AI's positive applications. Safety issues should be solved iteratively as the technology evolves. Stopping innovation is not the way to protect society; instead, we should solve safety problems as they arise, fostering a solution-oriented culture around AI.

Conclusion: encouraging investment and adjusting regulation

AI holds the potential to revolutionise industries, improve productivity, and create new markets that will enhance prosperity. But this future will only be realised if we adopt a pro-innovation approach that encourages investment, research, and new business models. By fostering a regulatory environment that is adaptive rather than restrictive, sectoral not horizontal, we can ensure that AI's full potential is realised.

Governments should encourage creative destruction through AI by supporting new firms and new models, rather than just reinforcing the dominance of existing players. AI’s promise lies not only in making current systems more efficient but also in creating entirely new opportunities that we have yet to imagine. If we choose to embrace this technological revolution, we will unlock growth and prosperity on an unprecedented scale. The AI revolution is here, and now is the time to seize its potential.