Logo and page links

Main menu


Finterai is such a new company, that it was not even registered when they started writing the application for the sandbox. They want to solve an issue far larger players have troubled with for years: money laundering and terrorist financing. Banks are required to do their share to prevent this, but are struggling to do so in an efficent way.

finterai 300x300.jpg– Banks spend 20 times more on compliance with the regulations for anti-money laundering than Økokrim manages to recover from money laundering, says Alexander Skage, CTO of Finterai. Nevertheless, DNB was fined NOK 400 million this spring for significant shortcomings in its anti-money laundering efforts.

Finterai will cut the financial institutions' risk and costs in the work against money laundering and terrorist financing. They will use federated learning, a decentralized AI method that is considered more privacy-friendly. With federated learning institutions might learn from each other's data without actually sharing it.

– We stand on the shoulders of giants, says Skage, and points out that they have gathered experience from similar pilot projects in Japan and the Netherlands. So he is confident that the method, which has not been tested to a great extent, will work.

The sandbox process

In the sandbox, the project will address two key issues:

  1. Is federated learning suitable for privacy concerns?

  2. Can federated learning facilitate machine learning across actors, in a way where personal information is not shared?

In the sandbox project, the Norwegian DPA will evaluate legal and technological issues in Finterai's solution, related to the privacy regulations. These assessments will be summarized in an exit report made publicly available.

Clarifications in these areas will not only benefit Finterai. Using federated learning, as a method of achieving machine learning without data sharing, will be relevant for the development of similar solutions in other areas. It will also have a lot to say for those considering investing in the development of federated learning systems, as this project will shed light on some important issues related to the technology.