Italian banks under pressure: AI accelerates the race for cyber security plans

Anthropic's warning about AI, which could control the internet, adds urgency to the work underway in recent weeks in the boards of directors and technical structures of Italian banks, for the creation of the examination and action plan on cyber risks linked to artificial intelligence. The request had arrived in recent months from the national authorities (Bank of Italy-IVASS) and from the ECB.

For the largest banks (significant) the end of year deadline has been brought forward to 31 October. The rapid development of AI, and the new Anthopic Mythos model, raised the alarm for surveillance last spring given that it is capable of circumventing current safety standards. A danger signal in addition to that of Anthropic in these hours on the possible control of the internet in 6-12 months. The plan requested from the credit institutions by the authorities will however only be a starting point and the bank's bodies, first and foremost the board of directors, will from now on have to consider increasingly rapid responses to digital risks and threats. Data from the Bank of Italy and experts show an increase in cyber attacks on the financial system in recent months.

And governor Fabio Panetta last May recalled how in 2023-2025 "incidents involving Italian banks increased by 80%. A picture that is changing with artificial intelligence, with "advanced models capable of identifying vulnerabilities in IT systems with unprecedented speed and depth". For this reason, the Bank of Italy has established contacts "with the main companies in the sector" and has started discussions "with national authorities, financial operators and their suppliers". As he explains to ANSA Tommaso Petrillo, Accenture risk manager for Italy and Greece "Artificial intelligence is changing not only the way in which banks innovate, but above all the nature and speed of the risks to which they are exposed. More advanced AI models can significantly reduce the time between the identification of a vulnerability and its possible exploitation, increasing the scale of threats and reducing the margin to react."

"For banks, this already translates into concrete actions: assessing their exposure, verifying the adequacy of safeguards, identifying and prioritizing gaps and defining an intervention plan with responsibilities, timescales and investments." A recent global survey conducted by Accenture on 775 risk professionals in 17 countries highlights how four out of five respondents believe that complex and interconnected risks are emerging more rapidly than ever; 66% also consider them more difficult to identify and manage." For the Accenture manager, "If AI accelerates threats, the way in which financial intermediaries govern risk must also change."