G7 Finance Leaders Agree AI Cyber Attack Tabletop Exercises

Finance ministers and central bank governors from the G7 group of advanced economies have agreed to hold tabletop exercises to prepare for cyber attacks driven by artificial intelligence, Japan’s finance minister Satsuki Katayama said on Thursday. No date for the exercises has been set.

Reuters reported the agreement from Bangkok, where the G7 finance leaders met on the sidelines of the International Monetary Fund’s annual meetings.

What was agreed

“The G7 finance deputies had been discussing the idea of doing tabletop exercises on frontier AI, which was signed off at the ministerial level,” Katayama told reporters, according to Reuters. “Frontier AI” usually refers to the most advanced AI models.

A tabletop exercise is a planned simulation in which officials walk through how they would respond to a crisis, such as a major attack on a bank’s systems, without any real systems being touched. It is a common way to find gaps in emergency plans.

A Japanese finance ministry official who attended the meeting gave more detail, Reuters reported:

Who runs them: the simulations will be developed by G7 finance authorities to test how they can respond to AI-driven cyber attacks.
What they are for: the results will test how well existing protocols work and will be used to revise manuals for handling such incidents.
Who takes part: the exercises, run by G7 finance ministries and central banks, could include private financial institutions “depending on the nature of simulations,” the official said.

Reuters did not report a schedule, names for the exercises or a list of participating institutions, and none has been announced.

Why now

The G7 agreement follows a warning from the IMF this week that confidence could be undermined by an AI-enabled cyber attack on a critical service of a financial institution, according to Reuters. The IMF’s October Global Financial Stability Report named cyber attacks as one of several near-term risks to financial stability, as TSN reported.

Regulators elsewhere are also moving. Reuters reported that a Bank of England deputy governor said on Thursday that the central bank is working with banks on deep cyber-recovery options, and that they can no longer assume their defences are immune to AI-enabled cyber attacks.

Reuters also reported that banking regulators worldwide are stepping up efforts to contain such risks, with some pointing to the potential for cyber attacks on banks or rogue agents to undermine global financial stability.

How it fits with other work

International bodies have been building rules for AI in finance. The Financial Stability Board recently finalised 12 sound practices for how financial firms adopt AI, and its chair, Andrew Bailey, warned the G20 about AI-related risks. The G7 exercises add a practical test of how authorities would respond if an attack happened.

What’s still unknown

There is no date, format or list of participants for the exercises. It is not yet clear which private banks or other firms might take part, or whether the results will be published.

What to watch

Watch for a G7 statement setting out a timetable, for announcements on private-sector involvement and for any revised crisis manuals after the first simulation.

What this means

Preparing for AI attacks. G7 finance authorities will rehearse responses to AI-driven cyber attacks.
Banks may join. Private financial institutions could take part, depending on the scenario.
No timetable yet. No date has been set for the exercises.