Selfie identity verification: A key part of fraud prevention

Countless Americans have fallen victim to identity theft at some point in their lives. In 2021 alone, 40 million saw their identity stolen, with the most likely victims ranging from 30 to 39 years of age—a key user demographic for digital businesses.

Moreover, studies show that as many as 87% of people have unknowingly allowed their personal information to be exposed online. This means preventative measures have become more important than ever, not only to bolster consumer trust and security but also to save money: It’s estimated that identity theft cost US businesses a total of $56 billion in 2021.

Using live selfie ID verification via mobile devices can help financial companies mitigate losses from theft while protecting their customers’ identities.

Sources: ProofPoint, McAfee, IDX

As technology and cyber security measures evolve, so do ways to skirt them. Synthetic identity fraud, for example, has become the fastest-growing financial crime in the United States. It works by combining both authentic and fake information to create a new—or “synthetic”—identity. The latter has enough verifiable information to seem credible, allowing it to be used to open fake accounts, make fraudulent purchases, and defraud a diverse array of businesses and organizations.

Even more insidious by nature, presentation attacks—commonly known as spoofs—are also on the rise. This form of identity fraud sees a bad actor use another person’s physical characteristics through things like photo printouts or fake fingerprints to impersonate them during an onboarding process or gain access to their existing accounts.

Biometric verification, including selfie identity verification, has thus emerged as a key method to combat spoofs by comparing a person's physical characteristics to verify they are who they claim.

What is selfie ID verification, aka real-time ID verification?

Selfie ID verification is a form of ID verification that requires the user to take and submit a selfie during the onboarding process. It then compares that selfie with the driver's license, passport, or other government-issued ID provided, in an action known as one-to-one verification. Generally, this is carried out with a smartphone, ensuring reduced friction for the user, and is required for highly secure services, such as opening a bank account or accessing sensitive information.

This process was once carried out manually, draining resources, limiting scalability, and lengthening onboarding times. Modern automated solutions like Plaid Identity Verification have changed all this, helping organizations break free from the above limitations while effectively mitigating identity fraud risk. They do so through facial mapping technology that compares specific facial features from the selfie—such as the distance between the eyes, nose, and ears—against those of the ID photo, ensuring they’re the same person.

Maximizing the effectiveness of selfie verification

To provide the highest level of security in a remote identity verification process, selfie verification should be combined with what’s known as biometric liveness detection, or a liveness check. This ensures that the user is physically present by requiring several selfie photos or videos to determine if they’re a real person or a passive fake. 

Also known as “anti-spoofing”, liveness checks cover a range of techniques used by authenticators to confirm the input being read is a true biometric source (e.g., an actual eye, thumbprint, or—in this case—human face) rather than a false or recreated image of one (e.g., a printed photo). These techniques look at a multitude of elements like skin reflectivity, patterns seen on screens, pixel changes and signs of image manipulation, depth, and more. 

Plaid Identity Verification, for instance, asks users to record short videos, rather than simply uploading a photo—though those with slow or unreliable internet connections may fall back to the latter (this, however, carries a higher risk of fraud as static photos can be more easily spoofed than videos). The verification then applies complex machine learning algorithms to verify the person’s authenticity, as well as face-matching algorithms to check the person’s identity against their photo ID. 

How reliable is selfie identity verification?

Statistics show excellent results when using selfie verification to scale. Plaid Identity Verification, for example, has achieved a success rate, also known as a pass rate, of up to a 94%. However, some selfie verification systems have failed to identify users properly, locking them out of the onboarding process entirely. More generally, facial recognition technology has also been shown to create biases that disproportionately affect people of color. 

This is because the decision-making capabilities of AI algorithms are ultimately based on training data, which can include biased human decisions or reflect historical or social inequities due to factors like:

• The size of the AI training base

• The source of the data used to create the training data sets and the level of real-world representation

• The labeling of the data sets 

• The quality controls that govern the tagging process

• The diversity of the developer team

These concerns should be taken into account when creating a machine-learning model. In addition, industry-specific testing should be carried out to help eliminate potential bias. 

The importance of selfie checks for AML compliance

Beyond mitigating identity fraud and its associated financial costs, the use of selfie verification can also help businesses maintain the highest standards of customer due diligence in their Anti-Money Laundering (AML) requirements. A key piece of these regulations includes verifying that a customer is who they claim. By staying compliant, businesses avoid the risk of potentially hefty fines and penalties.

Plaid Identity Verification achieves this through a multi-pronged approach that: 

• Matches personally identifiable information (PII) such as name, date of birth, phone number, and social security number against authoritative data sources that are governed by regulatory rules (think financial institutions or phone records)

• Verifies the authenticity of the government-issued ID document provided, ensuring it’s not a printout or has not been digitally edited

• Confirms the user is who they claim through automated selfie verification technology and liveness checks.

→ Want to fight fraud while handling AML requirements? Plaid Identity Verification is the lowest friction identity verification experience available.

Robust protection for today’s evolving threats

With identity theft growing ever more advanced—and in-person identity checks highly onerous—sophisticated digital identity verification has become an essential tool in mitigating fraud and carrying out customer due diligence. 

Scalable, low-friction, and highly reliable selfie ID verification provides a high layer of security for sensitive online operations without sacrificing the equally important user experience—ideal for maximizing the growth of your business.

Learn more about Plaid Identity Verification