Academics warn against overuse of machine learning

Lack of data makes AI technology unsuitable for risk management, say Cont and Rebonato

Backward-looking nature of machine learning could restrict its use in risk management

Market participants are increasingly exploring how machine learning can boost everything from credit underwriting to derivatives margin optimisation, but two prominent academics have warned that the technology’s applications for risk management are limited.

Speaking during a panel discussion at Risk’s Quant Europe 2017 conference in London on March 14, Rama Cont, professor of mathematics and chair of mathematical finance at Imperial College London, said a lack of data and constantly changing