Fridge hacks and finance attacks flag AI insecurity: Mike Dolan
By Mike Dolan
LONDON, Nov 3 (Reuters) - Hacking supermarket refrigerators may sound trivial. It's not. It's a warning of the economic disruption AI could wreak in an increasingly vulnerable world.
In the past few days alone, unknown criminal groups or state actors have been responsible for the biggest cyber theft in history, the hacking of Ukraine's air defense alert systems and the overnight shutdown of supermarket freezers across Europe.
https://www.reuters.com/graphics/EUROPE-CYBERATTACK/akvelwndxpr/chart.png
Attempted breaches of encrypted financial accounts along with economic and military sabotage have become almost routine. Some industry estimates put the global cost at trillions of dollars as financial sector targetsincrease.
If this activity is being supercharged by models and systems that even leading AI labs fear may undermine human systems and humanity itself, economic costs are bound to rise. Investment safety may need rethinking, and market volatility could be stoked.
Fears about risks to digital encryption in the banking and investment world, for example, have typically been reserved for the possible advent of usable, scalable "quantum" computing — still many years away in the view of many experts.
But the speed with which AI can already cause economic and financial havoc, whether in criminal or state hands, or even through rogue AI models, requires assessment.
https://www.reuters.com/graphics/BRV-BRV/egpbgxqoevq/chart.png
Much of the financial market obsession with AI has focused on whether the historic investment boom in AI infrastructure can continue at this pace, and on the payoff for firms spending hundreds of billions to scale it and for frontier AI laboratories.
Last month's severe tech-stock shakeout came amid doubts about the latter, as financing tightened and so-called hyperscalers burned cash. That, in turn, clouded the former, as doubts about eventual profits risk slowing the buildout.
Yet a rapid acceleration in cyber risks doesn't seem top of mind in the macro world. That's partly because hacking has been seen predominantly as a micro issue affecting individual firms.
In a period of rising geopolitical stress, however, state-backed cyber operations appear to be moving beyond espionage and into attacks on public infrastructure, security, supply chains and finance. They could well be reaching a different level.
https://www.reuters.com/graphics/USA-CYBERINSURANCE/AI/klpyoyowrpg/chart.png
GEO-HACKING
Military leaders around the world assume they are now in an arms race over cyber and AI-related capabilities, with the use of both in the ongoing Iran war trumpeted as a major factor, not to mention Ukraine's advanced drone war capabilities in fending off Russia's invasion for four years.
Rising defence budgets in the US, Europe and Asia will increasingly be spent as much on these areas as on tanks, ships and planes.
A monitor compiled by the Council on Foreign Relations showed that, over the past 20 years, some 77% of operations were suspected to have been sponsored by China, Russia, Iran and North Korea.
Business leaders have also flagged rising anxiety. Earlier this year, the World Economic Forum's report on dominant risks to the global economy polled its members and experts and showed cyber worries were both high and rising.
The list of most pressing current risks included "misinformation and disinformation," "adverse outcomes of AI technology" and "cyber insecurity," all in the top 10.
Over a two-year horizon, "misinformation and disinformation" and "cyber insecurity" ranked as the second- and sixth-biggest risks, respectively. The largest rise in concern on a 10-year horizon was the 25-place climb in "adverse outcomes of AI" to the No. 5 spot.
There's also some indication that private investors are starting to chime with that picture.
A survey by Britain's Association of Investment Companies (AIC) on Monday showed considerable wariness among members about AI's overall impact, with almost three-quarters of respondents worried about "misinformation," ahead of job losses, erosion of human skills, privacy concerns and environmental pressures.
"As long as the AI investment boom continues and investors are making money, such fears may be overlooked," said Nick Britton, research director at the AIC. "But if there is a market correction then AI companies may have to work harder to persuade investors that they can deliver growth while addressing their wider concerns."
In the apparent absence of a ransom, hacking fridges may seem like wanton vandalism. But the willingness, super-sized capability and political scope of such actions reflect a more dangerous and unpredictable global environment, one that may come with a hefty price tag.
(The opinions expressed here are those of Mike Dolan, a columnist for Reuters.)
Enjoying this column? Check out Reuters Open Interest (ROI), your essential new source for global financial commentary.
Follow ROI on LinkedIn, and X.
And listen to the Morning Bid daily podcast on Apple, Spotify, or the Reuters app. Subscribe to hear Reuters journalists discuss the biggest news in markets and finance seven days a week.