In a private talk that I gave in early September, I pointed out that optimism about artificial intelligence ( AI ) is widespread in the financial and high-tech worlds, but almost nowhere else. This struck me as odd, because those in finance should know that rising interest rates – especially longer-term rates – will complicate things. Just as tighter financial conditions have tested the scale and depth of the market’s optimism in past investment cycles, so might they do so again. Thus, I suggested that the underlying inflation outlook would be among the most important indicators to watch.
Most of those in the room looked at me like I was some sort of neanderthal who just doesn’t “get it”. Apparently, I have failed to realize that AI is an unstoppable force. Still, I pointed out – for good measure – that investors also seem unfazed by rising fears of mass joblessness or rogue AI systems disrupting critical infrastructure or killing us all. Again, the financial world’s complacency struck me as odd, given that a public backlash could also complicate matters for the AI bulls.
The American public made its position known on November 3, decisively rejecting the incumbent Republican Party, which many associate with the AI and tech industry, and overwhelmingly voting for restrictions on data centres almost everywhere that they were on the ballot. A day after markets absorbed these results, we witnessed a large-scale sell-off in AI stocks. Is a more sober outlook now setting in?
To be sure, AI could significantly boost productivity in key sectors of the economy. But only if it is harnessed correctly ( as I argued over a year ago ). It already can help employers replace workers for some functions, which may boost productivity in a strictly technical sense. But unless we are extremely confident that AI will drive productivity higher, raise real ( inflation-adjusted ) wages and spread the benefits widely, then what exactly is the point?
Why would we want a technology that simply dashes people’s hopes of finding a fulfilling economic purpose, or that threatens to disrupt critical infrastructure or kill us all? More to the point: Why should investors want these outcomes to be part of the narrative? With dire warnings already in wide circulation, it has been obvious that society will demand a policy response from elected leaders. The US midterms offered merely a preview of what is to come.
The focus in recent months has been on data centres and debates about NIMBYism ( “not in my backyard” ). But the issue of regulation will also now be on the table, despite whatever resistance the next Congress, firmly under Democratic control, faces from US President Donald Trump.
That is no bad thing. As the authors of a forthcoming paper for Brics+ Thinking ( a policy platform I helped establish earlier this year ) argue, some kind of global AI regulation is imperative, regardless of how difficult it may be to negotiate. Doing so will require coordination not only between the United States and China, but also with the so-called middle powers and the Global South.
Those in finance should remember that, since the 2008 crisis, many parts of the world have endured a long period of subdued economic growth. And although the US has grown more than most European countries and the rest of the developed world, it, too, has fallen beneath what my Goldman Sachs colleagues and I expected for it when we were forecasting at the turn of the century.
Much of this subdued growth has been due to weak productivity and, until recently, anaemic private sector investment. We have come through a sustained era of low real wage growth alongside strong profit growth and record-breaking equity valuations.
If AI simply supercharges these market dynamics and fails to deliver a much wider dispersion of genuine productivity gains, a dramatic upheaval will become all but inevitable. Perhaps the historic US midterm results will be heeded as a warning, leading to a change in narrative and in the strategies determining how these technologies are designed and deployed.
It would be a shame if such a powerful technology failed to bring benefits to many segments of economic and public life. As I have noted previously, a lack of productivity growth helps to explain many of the challenges now facing the United Kingdom’s National Health Service, and it is just one of many examples.
Optimists may insist that everything will turn out alright in the end. With public pressure rising in so many countries, and with the technology holding so much potential, perhaps it will be directed towards powering widely shared growth. But simply wishing for this outcome won’t be enough.
Jim O’Neill is a former UK Treasury minister and chairman of Goldman Sachs Asset Management.
Copyright: Project Syndicate
