Confidential Computing: Introducing New Capabilities in the Fight Against Digital Fraud

Fraudsters thrive in times of uncertainty, and the pandemic proved to be their perfect foil. Financial institutions had to accelerate their plans for digital transformation and the rush to implement new technologies and processes led to blind spots that were taken full advantage of: 

• Approximately $80 billion was fraudulently accessed from the Small Business Administration’s Paycheck Protection Program. That amount pales in comparison however to the $90 to $400 billion (a closer estimate is still in development) stolen from the Covid-19 unemployment relief fund.¹
• The Federal Trade Commission announced another record year for consumer-reported cases of identity theft in 2021, with 1.4 million incidents.² 
• Suspicious Activity Reports filed by depository institutions reached a new record in 2021, increasing 21% over 2020.³ 

This swell in fraudulent activity is enabled by increasingly sophisticated approaches. In some cases, criminals use artificial intelligence to bypass legacy fraud detection solutions and leverage automation to operate at a massive scale. One thing is clear – new approaches are needed to help financial institutions fight back against digital fraud. 

 Why Confidential Computing is Set to Change the Game 

Financial institutions have been slow to adopt cloud technology for a number of reasons: the complexity of existing legacy solutions, skill gaps in the workforce, and fragmented compliance requirements. But perhaps the greatest fear that SaaS solutions present for financial services companies is the perceived lack of control and sensitivity of the data. 

For financial institutions hesitant to share sensitive customer data on the cloud, Confidential Computing provides secure access to aggregated fraud detection models that on-premise solutions simply can’t provide. That’s why we’re proud to be working with Intel to bring Confidential Computing to financial services. 

Protecting Sensitive Data with Enclaves 

While existing solutions secure data when it’s sent across networks, Confidential Computing also protects data while it’s being stored and used. It does this by running applications and data in a form of impenetrable hardware memory, called an “enclave”. 

Even if a system is compromised, any data or code that is isolated in an enclave is protected. In effect, this means that Confidential Computing protects data from the most sophisticated hacking attempts, as well as more straightforward security concerns, such as stolen hardware. 

Adding a Further Layer of Security with Attestation 

While running data in an enclave offers a high level of security, some applications of Confidential Computing go one step further by applying attestation.  

This is a process that verifies the security of an enclave, the application being run in it, and the data that’s being processed.  

For example, the Intel® Software Guard Extensions (Intel® SGX) attestation process helps an enclave to prove that: 

1. The code built and validated by the user is running unmodified in a genuine enclave. 
2. The hardware in which it is running is a secure platform with all the necessary updates. 
3. The hardware and software configurations needed for the enclave are correctly applied. 

Confidential Computing Powered by Intel® SGX  

FiVerity’s platform is powered by Intel® SGX to provide financial institutions with greater protection of sensitive customer data and improved access controls.  

Intel® SGX enables data to be processed in memory without exposing it to the rest of the system, dramatically increasing the security of sensitive data in use. With Intel SGX, encryption keys are protected as well, both at rest and in use, for added security when data is being processed. 

How Confidential Computing Powers Fraud-Fighting Tools 

Now that we understand the technical benefits of Confidential Computing, how can financial institutions apply this technology? 

Multi-Party Computing 

Confidential Computing gives financial institutions new tools for the fight against digital fraud through secure access to aggregated fraud detection models. 

Financial institutions that had previously limited themselves to on-premise technology due to security concerns can confidently embrace cloud sharing and, in doing so, gain access to collaborative learning models that aid the fight against digital fraud. 

These models analyze encrypted data from multiple banks to identify patterns of fraudulent activity with a much higher degree of accuracy. Information about fraudulent activity from each financial institution is provided anonymously, ensuring other users aren’t able to attribute data to a specific company.  

This allows users to benefit from access to a much larger pool of information about fraudulent activity without the fear of exposing their sensitive data or providing competitors with insight into the fraud occurring within their portfolio. 

Disadvantages of On-Premise Solutions 

Despite Confidential Computing offering financial institutions secure access to cloud-based models, many organizations continue to limit their fraud detection efforts to on-premise solutions that provide a limited view of the fraud landscape. 

One downside of on-premise models is that they quickly suffer from degradation, becoming stale shortly after they’re implemented. 

Another is the difficulty of keeping up with changes in the data and accurately detecting emerging patterns requires ongoing attention on a larger scale. For example, a bank receiving a high volume of applications from recent Ukraine immigrants is an unanticipated event that an internal model may not be equipped to interpret, potentially creating a slew of false positives. 

Cloud-based solutions ensure banks have access to the latest models and a larger data set that helps to accurately detect fraudulent activity. 

Now is the Time to Embrace Confidential Computing  

With digital fraud growing exponentially during the pandemic, and continuing to be a thorn in the side for financial institutions, new fraud detection solutions, powered by Confidential Computing, could not have come at a better time. 

Digital fraud is a serious threat, but legitimate consumers and businesses are increasingly eager for the convenience and cost that online banking solutions provide. Financial institutions now have the tools they need to quickly identify fraudulent attempts and focus their efforts on assisting these legitimate customers.

1. NBC News, Mar 28, 2022. “Biggest Fraud in a Generation: The Looting of the Covid Relief Program Known as PPP.
2. FTC, Mar 2022. Consumer Sentinel Network Data Book 2021. 
3. U.S. Treasury Financial Crimes Enforcement Network, 2014-2021. Suspicious Activity Report Statistics. Depository institutions only.