Complex Data Risks with the Adoption of AI by Boyke Baboelal
A featured article of our January 2020 edition of PRMIA's Intelligent Risk quarterly newsletter
At the same time as regulators are focusing attention on data quality and data risk management processes, the complexity involved in managing data risks has become greater than ever.
The growth in data volumes needed for regulations and capital rules such as MiFID and FRTB, as well as the additional complexity in processing that data, such as mapping transaction data to risk factors and proxying non-modelable risk factors, has resulted in increased organizational exposure to incomplete, inconsistent or inaccurate data. However, these data risks are small when compared to the data risks coming from the adoption of AI. New risks such as the use of biased data, poisoned data (malicious data injected in the model), and dirty data can have severe legal and reputational ramifications.
Continue to Read [PDF]
About the Intelligent Risk
is PRMIA's quarterly publication, bringing all PRMIA members free access to knowledge and information about risk management for financial institutions as well as current information on PRMIA chapters, committees, academic partners, news and events.
Individual articles from each edition are published under our members only Risk Library
resources section. PRMIA is sharing select articles from the January 2020 edition with the public. Get more articles like this by joining PRMIA today.
PRMIA Membership Information