MILAN (Reuters) – Italian market watchdog Consob has been testing artificial intelligence (AI) for use in the initial prospect listing approval process and to help identify insider trading, the authority said on Tuesday.
“These are algorithms that could support preliminary analyzes for the detection of suspected cases of insider trading, to be followed by targeted investigative activities,” the watchdog said.
“The algorithm takes three seconds to detect an error, the expert human analyst takes at least 20 minutes,” the official said in slides attached to its annual report.
Testing activities were carried out last year on two prototypes developed in-house in 2022 together with the Scuola Normale Superiore university in Pisa, and a third model was also developed.
The adoption of AI should improve the quality of controls to detect irregularities. The next step will be to move from the prototype phase to incorporating AI into the market regulator’s regular activities.
Regulators and watchdogs in other jurisdictions have been interested in exploring the potential of AI.
In Britain, the Financial Conduct Authority has been using AI to help protect consumers against online fraud.
(Reporting by Romolo Tosiani and Giancarlo Navach; Editing by Keith Weir)