FCA reports that synthetic data can enhance AML innovation and collaboration by enabling safer data sharing and more effective testing of financial crime detection models

The project assessed how synthetic datasets could support the development and validation of AML tools without exposing sensitive customer information, helping firms overcome barriers linked to privacy, confidentiality, and restricted access to real transaction data. Findings also highlight limitations relating to data fidelity, model bias, and governance, with the report recommending continued industry collaboration and technical development to ensure synthetic data remains reliable and effective for AML/CFT applications.