R&D / Banking AI Evaluation Library / Datasets and sources
AMLworld (IBM synthetic AML transactions)
Synthetic transaction data sets in several sizes, where each transaction is labelled as laundering or not. The paper calls them IT-AML.
Open the publisher page: IBM Transactions for Anti Money Laundering (Kaggle) (opens in a new tab)
Dodone links to the publisher and does not host or copy this data. Read the licence on the publisher page before you use it.
About this source
- Publisher
- IBM Research (E. Altman and others)
- Context
- Simulated financial transactions with money-laundering patterns
- Version
- Not available (Kaggle shows the last update as about a year ago)
- Last checked
- 2026-11-03
- Labels or answers
- A laundering label for each transaction, from the simulator.
Facts stated by the publisher
- The paper describes synthetic data sets in small, medium and large sizes.
- The largest low-illicit set has about 176 million transactions.
Licence, access and privacy
- Licence
- Community Data License Agreement – Sharing – Version 1.0 (opens in a new tab) (Licence verified)
- Scope
- The Kaggle page states this licence. It has sharing terms for anyone who publishes the data or changes to it.
- Access
- Registration needed. Downloads from Kaggle need a Kaggle account.
- Privacy
- Privacy reviewed: low risk. Fully synthetic data from a simulator.
Limitations
- A transaction label is not a full investigation outcome.
- Simulated laundering patterns may differ from what a given bank sees.
Suggested for these evaluations
- Financial-crime alert analysis Diagnostic only
Why it helps: Synthetic transactions with laundering labels test pattern detection and arithmetic.
What it does not validate: A full investigation, the right next step and the effect on customers.