bunq amps up security after scam claims; partners with NVIDIA to combat financial frauds using GenAI

Amsterdam-based neobank bunq has been the focus of numerous media publications, with concerns raised about the app’s security and customer service.

According to NVB, in 2023, the Netherlands saw as many as 10,000 cases of bank helpdesk fraud and thousands more phishing victims, mentions bunq in its mail for its account holders

At bunq, there were 85 in total.

“Online fraud is a huge problem. Not a bunq or banking problem, but a social problem. A problem that we can only solve together if we dare to look at the objective facts,” says Ali Niknam, founder and CEO of bung in a LinkedIn post. 

As a result, the Amsterdam-based neobank has announced some of the significant changes to amp up security.

“Because our app has maximum security, criminals choose to deceive you. For example, by posing as a bunq employee and calling you. They try to obtain your login details (and biometric data), ask you to install software, or convince you to transfer money yourself. To combat this type of fraud, vigilance is the best medicine. That is why we are working hard on numerous improvements to increase awareness of this type of fraud,” says the company.

Some of the features include:

Deferred payments: With deferred payments, users will have the option to delay payment for an additional 24 hours, allowing them the opportunity to cancel the transaction if necessary.

This feature is made possible through the combination of various security systems that work together to determine whether a payment should be delayed.

Additionally, bunq’s AI system is constantly monitoring transactions to prioritise user safety.

bunq also said it has noticed that some individuals transfer money to scammers using their name. 

In response, they have added extra warnings if the name doesn’t match the IBAN, providing the option to cancel or confirm the payment.

Safety Lock: If bunq suspects that there may be an issue with your account, all payments and passes will be paused, and users’ access to bunq will be frozen. To resume access, they will need to confirm an additional face scan along with your login code.

Third-Party Access to Your Device: Sometimes, individuals are persuaded to install software that allows third parties to access their computer or phone. AnyDesk is a well-known example of such software.

The scammer then has access to your entire system, such as all your emails, your files, and thus your banking environment. Using new security, bunq tries to detect if this is the case. And if so, the neobank will restrict the use of your account.

“We are busy implementing improvements and will devote even more time and attention to this in the near future. Even before the recent media attention, we were in direct contact with victims; For many, we are still trying to get money back, for others, we are looking at the extent to which a leniency fee is appropriate,” adds Niknam.

bunq joining hands with NVIDIA

bunq is combating financial fraudsters with the help of NVIDIA accelerated computing and AI.

bunq turned to generative AI to help detect fraud and money laundering.

Its automated transaction-monitoring system, powered by NVIDIA accelerated computing, greatly improved its training speed.

“AI has enormous potential to help humanity in so many ways, and this is a great example of how human intelligence can be coupled with AI,” said Ali el Hassouni, head of data and AI at bunq.

Bunq’s AI-powered system uses supervised and unsupervised learning for automation and scalability.

NVIDIA GPUs accelerated bunq’s data processing by over 5x, and the open-source NVIDIA RAPIDS suite enabled nearly 100x faster fraud-detection model training, claims the company.

RAPIDS is part of the NVIDIA AI Enterprise platform, streamlining data science and AI application development and deployment.

“We chose NVIDIA’s advanced, GPU-optimised software, as it enables us to use larger datasets and speed the training of new models — sometimes by an order of magnitude — resulting in improved model accuracy and reduced false positives,” says el Hassouni.

The company is exploring NVIDIA NeMo Retriever, a collection of generative AI microservices available in early access, to further improve Finn’s, bunq’s personal AI assistant, accuracy.

“Our initial testing of NeMo Retriever embedding NIM has been extremely positive, and our collaboration with NVIDIA on LLMs is poised to help us to take Finn to the next level and enhance customer experience,” says el Hassouni. 

Note: Ali Niknam’s quotes were translated using Google’s translation