What is AML AI?
AML AI is an API-based machine learning pipeline for automatically training, testing, deploying, and monitoring a productionized anti-money laundering (AML) model. As a managed service, Google takes care of the infrastructure behind the scenes and presents teams with a production-ready system to train, predict, and backtest models to tackle money laundering.
Interface
The main way of interacting with AML AI API is using the https://financialservices.googleapis.com endpoint with REST API calls. The Google Cloud CLI tool is not supported for directly calling the AML AI API, but it is recommended to use the Google Cloud CLI to obtain credentials.
You may want to use programming languages to interact with AML AI. To make coding against AML AI easier, Google provides generic API client libraries for a number of different languages that can reduce the amount of code you need to write and make your code more robust.
Each of the API client libraries provide a means to use application default credentials (ADC).
For details about the REST interface, see Financial Services API.
Data
AML AI reads input data from BigQuery and writes output predictions and backtest data to BigQuery. For input data, an AML AI dataset resource must be created which references the data in BigQuery. This dataset must reside in the same location as the AML AI instance.
The AML AI dataset resource represents pointers to datasets in BigQuery. It does not hold or point to any specific snapshot of the data in these tables. If data is modified after a dataset is created (for example, if records are deleted), this will be reflected in the results of other calls to the API (for example, the creation of new models or when running predictions). Modifying the data this way is not recommended. For more information, see Create and manage datasets.
Services used by AML AI
As well as the AML AI API itself, there are a number of other Google Cloud API services which are required to use the AML AI:
Required
- Cloud IAM: for identity management and access management
- Cloud KMS: for key management
- BigQuery: for data storage
- Cloud Logging: for logging and monitoring
Optional
- Cloud HSM: Optional hardware-backed storage for encryption keys
- VPC Service Controls: Prevent data exfiltration to unauthorized networks and devices
