Five hot topics in online fraud and identity verification for 2024

1. Reusable, user-centric identities

Also referred to as portable identity, reusable identity obviates the need to constantly enter and re-enter your credentials as you perform different tasks online. Instead, a single digital identity represents you across the web. Studies show the average person has around 100 online accounts, so reusable identities could save individuals a lot of time and stress. In their decentralized, user-centric form, they also offer far more control. 

"I can see who has access to my digital data, who’s looked at it, why did they access my data, so I have a lot of transparency in that sense," says Viktoria Ruubel, MD Digital Identity at Veriff.

Meanwhile, for companies, reusable identities reduce the cost of customer identification while minimizing the risk of reputational damage and financial liability.

"It has more power to protect from fraud, because you understand the maturity of the identity, you understand who owns it," explains Viktoria.

2. Improving the user experience

At Veriff we admit to being obsessed with all things identity related, but we understand that what matters to users about identity verification is what it allows them to do. Taras Boyko, Product Manager at Veriff, sees three key ways to make identity verification more user friendly in the future. 

• Optimizing the flow: Automating tasks like capturing images and video or identifying the document type presented with the help of artificial intelligence, thereby simplifying the process and reducing friction.
• Demystifying the experience: Explaining to users what data is being collected, why it’s needed and how it will be used, to build trust and help make identity verification an accepted part of everyday life.
• Creating unified systems: Developing a more collaborative approach between governments and specialist private companies like Veriff to reduce siloes and ensure more efficient and effective fraud prevention.

3. The importance of dynamic, agile solutions

Fraud prevention has always been a cat and mouse game between businesses and the criminals who want to defraud them and their customers. With the advent of artificial intelligence, the game is becoming ever more serious.

"A fraudster will basically probe your system until they figure out the rules, and then go around them," says David Divitt, Senior Director of Fraud Prevention and Experience at Veriff. "So, you need to have good tools in place that you can use to dynamically adapt your strategies to fraud trends as they evolve."

Artificial intelligence and machine learning can be a powerful asset in the battle against online fraud. Instead of a static set of rules waiting to be discovered, businesses can use AI to present a moving target. This is more and more necessary as fraudsters themselves are enthusiastically adopting AI.

"Investing in good tech to help fight the bad tech, is the way you should think about it," says David.

4. Reinforcement learning from human feedback

The impressive abilities of the latest generation of generative AI have led to concern in some quarters that the human workforce will soon be surplus to requirements. In reality, like computing before it, AI works best when it is treated as a useful tool rather than a direct replacement for humans. 

Reinforcement learning from human feedback (or RLHF for short) is an approach that uses human intelligence to review outputs from an automated process such as a machine-learning based model. Analysts identify and address issues, enriching, augmenting and even correcting the data before reinserting it into the process.

"That really is helpful because it serves as a fresh input to the model, so the model becomes that much smarter, and the output is that much better,” says Suvrat Joshi, Senior VP of Product at Veriff. "Essentially you inject humans into the loop in order to build a better model."

Identity verification is a prime candidate for RLHF, since the ability to stop fraud while maximizing conversions is highly dependent on the quality of the overall dataset.

5. The growth of age estimation

Online fraud prevention is increasingly based around identity verification using official documents such as a user’s passport or identity card and comparing this against live video of their face. This solution works extremely well for scenarios where accuracy is the priority and users have access to a suitable ID document. But what about situations where users want to access age-gated products and are either unwilling or unable to share official documents for ethical, privacy-related or practical reasons?

Veriff’s Age Estimation solution uses facial biometric analysis to estimate a user’s age in cases where no formal identity document is provided. This makes it perfect for providing access to age-gated content such as computer games, where young users may be old enough for the content to be age appropriate, but not old enough