And yet, I have reconsidered my own views. As an economist, in recent years I have believed that disruptive innovations in artificial intelligence must be a net positive, because that is what has happened with the emergence of new technologies throughout my career. Reality has never validated the arguments of technoskeptics and pessimists that some new technology will lead to a massive unemployment crisis, so my default assumption was that this time should be different. Moreover, AI could be a boon for Western societies facing challenges such as a rapidly aging workforce, political backlash against immigration, and a growing burden of public debt.
But in response, I often hear from those most deeply immersed in the field that I “don’t understand” how frighteningly fast AI is advancing. I have failed to appreciate how dangerous this technology could be, especially if it reaches the point where it can adapt its own thinking without any input from a controller. At the same time, if AI ends up in the wrong hands – whether in criminal organizations or rogue states – it could lead to political and social chaos.
In addition, there is the extreme cost of producing and operating this technology. We just watched US President Donald Trump’s visit to the UK, and along with all the pomp and ceremony, he paid close attention to mobilizing even more investment in data centers and the energy infrastructure to power them.
Given the unprecedented market valuations and the scale of the sums involved, I have to ask: what, exactly, is the goal? If AI, for all its obvious power, can’t be applied in a way that materially solves our biggest problems, what was it all for? I understand why the owners of leading companies want all other businesses to adopt AI. But that doesn’t mean the technology will help us solve the biggest and most pressing problems facing our society.
For private sector employers, AI seems like an obvious way to contain labor costs. Most economists welcome anything that can shock corporate executives into not obsessing over the need for cheap and flexible labor. But on the whole, this issue is not that important.
In the UK, for example, productivity growth has been weak since the 2008 financial crisis. And if you dig into the data, you’ll find that key public institutions are the main driver. And while everyone talks about improving productivity in the NHS, no one makes it a priority.
When I served on the Times’ independent commission on health in 2023, I often had to fight off the widespread view that, regardless of our recommendations, health spending as a share of GDP would inevitably continue to rise. But why should this be the case? In the late 1970s, health care spending as a share of GDP was about the same as education – about 5%. But today they are approaching 12% of GDP, more than twice as much as education spending.
In my economist’s view, these trends don’t make sense. If anything, education spending should be higher than health spending because better educated young people can both increase the productivity of the future labor force and make better health choices. It may also be easier to achieve higher productivity in health care than in teaching. Even worse, the more health spending rises, the harder it is to avoid spending cuts in other areas unless taxes are raised. The UK Office for Budget Responsibility continues to forecast appalling levels of potential national debt by 2050 if we do nothing about the NHS, welfare payments, state pensions and other social services.
So my recommendation to government politicians, industry leaders and others is to develop a serious, clear, evidence-based plan to ensure AI contributes to productivity growth where it is needed most. It won’t be easy, but given how much brainpower and talent has gone into developing this technology, surely someone will figure out how to put it to good use.
Jim O’Neill,
former UK finance minister and former chairman of Goldman Sachs Asset Management.
© Project Syndicate, 2025.
www.project-syndicate.org
