Engaging with civil society is the best approach to limit the risks posed by AI
Over the past year it’s been hard to avoid news stories about Artificial Intelligence (AI) and the impact it could have on our lives. As we approach the end of 2023, governments around the world are grappling with the best way to reduce the risks, and harness the benefits, presented by technology that has previously only existed in the realm of sci-fi. The UK has taken a central role in leading this debate, recently hosting the global AI Safety Summit to consider the risks of AI.
The debate on the impact of AI has been dominated by warnings of existential threats to humanity’s existence, international security and robots replacing jobs. These types of ‘robot apocalypse’ scenarios have been expressed by big names in the AI and technology world, the same people who stand to profit most from this technology, but who are also – perhaps curiously – calling for immediate action from governments to avoid catastrophic outcomes. It’s unusual to have a powerful industry call for regulation. On the one hand, this is a fitting reaction to a complex technology that’s developing at a significant pace with potentially unprecedented consequences.
On the other hand, we should perhaps question the motives or hidden agendas of some of these actors. Many of the voices calling for immediate regulation have come from big tech (such as Google and Microsoft) and other highly powerful and influential actors such as Elon Musk and Open AI (the developers behind ChatGPT). They have the financial and computing capability to capitalise on the benefits presented by AI to maintain and vastly increase their market dominance. They also have the lobbying power to influence the rulemaking process – and perhaps their calls for regulation are simply an attempt to get ahead of the curve, and shape the regulations to their advantage.
Experts have warned that governing AI effectively is a “sociotechnical challenge: one that requires the meaningful involvement of people and communities, particularly those most affected by the development and use of AI systems.” In no sector is this more important than in financial services, where the complexity of the system – and its impact on society and the wider economy – means that we need to hear from a wide range of perspectives and expertise to develop a wide range of solutions.
Given how enormous the consequences of AI could be, it is vital that the government and regulators actively engage with all parts of society to gain a fair and representative understanding of the potential impacts, and focus on supporting the overall public good. This will result in fairer policy outcomes and protect against the risk of regulatory capture by the powerful tech and financial services industries.
As we argue in our new report, the impacts of AI on the financial system are being ignored by the government. The financial system works well for those at the top, and very badly for those at the bottom. If it’s left to big tech and other powerful lobbies such as the big banks to decide how to use AI, then it will just supercharge the inequities of the current system and risk another financial crisis. To avoid repeating and exacerbating the mistakes of the past, governments need to make sure other voices can be heard, and positive social outcomes are prioritised over profit.
Earlier this year the government set out its intention to take a light-touch approach to regulation. Following the Summit, the government set-out agreements between world leaders and big tech firms to collaborate on testing the next generation of AI models against harms. More recently they confirmed that they will not regulate at all in the short term as this may stifle innovation.
We believe that this approach is short-sighted, undemocratic and potentially dangerous. A focus on innovation does not automatically result in better outcomes for consumers or society, and there is no guarantee that any economic or productivity benefits of AI-led innovation will be shared across society. Indeed, without active government intervention, it is far more likely that they will flow to those at the top of the economic pyramid, to the exclusion and detriment of those at the bottom. Further, the agreements do not address harms that are currently happening today or those affecting financial services.
In the lead-up to the Summit we signed a letter with 100 other organisations that warned the Prime Minister that the lack of civil society participation means that the voices of the communities and workers most affected by AI will be marginalised. Without this type of society-wide engagement, there is a risk that the development, introduction and monitoring of AI processes in financial services will not serve society.
With technology as powerful as AI, and an industry as influential as financial services, ensuring that the public interest comes first should be central to the government’s approach to AI.