Global identify verification boosts government compliance efforts and aids in the safer flow of commerce. Veriff has completed more than 100 million such verifications. The company said it can verify 12,000 documents from more than 230 countries and territories.
The AI Innovator chats with Veriff CEO Kaarel Kotkas to find out how AI fits into ID verification. The following is an edited transcript of that interview.
The AI Innovator: Veriff leverages AI for identity verification. How has AI transformed the accuracy and speed of this process compared to traditional methods?
Kaarel Kotkas: Veriff, a global identity verification and authentication platform, provides advanced technology that uniquely combines AI-powered automation with reinforced learning from human feedback, deep insights, and expertise. Our identity verification platform analyzes more than 1,000 data points, including the person, document, device, and network, to enable trust between businesses and their users.
We’re harnessing the power of artificial intelligence to improve both the user experience and the accuracy of its solution. At a time when fraud is on the rise globally, companies need to ensure they are protecting their business and their customers against threat actors.
Veriff’s advanced multi-layered identity authentication stack utilizes AI to more quickly identify fraud, from deepfakes to fake documents or IDs. Our technology learns from previous fraud attempts to flag common signs such as blurry ID images, suspicious backgrounds, and known threat actor IP addresses. Veriff’s artificial intelligence finds patterns and anomalies to better identify and stop traditional and emerging fraud attempts.
What are some of the opportunities and challenges you’ve faced while training your AI to detect fake or manipulated identification documents?
AI adoption across the globe is still in its early stages. The technology offers tremendous opportunity, but there is still a learning curve. We have implemented AI technology to keep pace with fraudsters who are also employing AI to create deepfakes and other emerging fraud tactics. When training our models, it’s still critical to have human oversight to prevent potential issues such as hallucinations or bias. All of our critical AI decisions are verified by a human expert. This enables us to responsibly grow our AI adoption, ensuring the best identity verification experience for our customers and their users.
How do you balance maintaining privacy for users while ensuring your AI models have sufficient data to learn and improve?
We are deeply committed to building a community empowered by trust. We train our models to spot anomalies in portraits, fonts, and characters, such as unintended text that signal fraudulent documents. Additionally, our systems cross-reference information like serial numbers, issue dates, and barcodes to ensure consistency. These methods help identify potential fraud by catching subtle irregularities that might otherwise go unnoticed. To maintain user privacy, Veriff follows the highest standards of privacy and data protection laws, such as the European Union’s General Data Protection Regulation (GDPR).
What role do AI-driven insights play in staying ahead of emerging fraud techniques, and how do you continually update your models to counteract new threats?
A recent survey found that nearly 78% of U.S. decision-makers have seen an increase in the use of AI in fraudulent attacks over the past year. On the flip side, nearly 79% of CEOs use AI and ML in fraud prevention. In times like these, we can’t hide from new technologies. Addressing the multifaceted and evolving nature of fraud requires a tailored, data-driven, ecosystem-based approach that relies on a multitude of tools.
Veriff uses deep-learning models built from in-house technology to grow, train, and deploy machine learning for any use case. Staying ahead of fraudsters requires collaboration between fraud operations teams and data scientists. This creates an ongoing circle of learning where the data teams identify patterns based on the fraud operations team’s input around new fraud techniques.
Blockchain technology is often touted as a way to enhance transparency and security. How do you see blockchain fitting into identity verification and financial services, especially when paired with AI?
Blockchain is designed for the secure sharing of information by creating an unalterable record of transactions. However, the anonymity and sheer speed of these transactions, like those with cryptocurrency, creates identity verification challenges. This is why criminals sometimes use cryptocurrencies in an attempt to evade conventional controls relating to money laundering.
Cryptocurrency needs to be made safer and more accessible, and that starts with eliminating fraudulent transactions. Identity verification in blockchain improves the safety of the user experience, helping financial institutions build credibility and operate seamlessly in diverse markets. The automation enabled by AI has the potential to help identity verification meet the speed demands of blockchain transactions.
Can blockchain help mitigate some of the risks associated with centralized identity verification systems?
Decentralized identity is one of the trending topics around Identity Access Management (IAM) in 2025. Decentralized identity allows users to control their identity data rather than relying on centralized authorities. Blockchain technology is expected to play a significant role in this shift.
For example, financial services use blockchain technology for identity verification in open banking. It can create decentralized, tamper-proof records of verified identities, making it easier to confirm user identity across multiple services.
How do you foresee regulation evolving around the intersection of AI, blockchain, and financial services?
No technology exists in a regulatory vacuum. Existing privacy, copyright, competition, anti-discrimination, and consumer protection laws apply and have always been applicable to AI and most automated solutions. Companies must consider these laws to form an enforceable set of existing guardrails, while ensuring they have the ability to prove compliance.
The financial services industry is already heavily regulated. And new technologies like blockchain and AI bring their own set of regulations. It’s important for companies to understand the intersection of these regulations and put systems and processes in place to ensure compliance and reporting, and the agility to evolve as these requirements change over time.
Technology providers like Veriif also need to comply and evolve with these regulations, such as changing how personal data is processed in light of privacy laws. We believe these laws can increase people’s trust in the industry and the technologies that support it.
What innovations or advancements in AI and machine learning do you think will most impact identity verification over the next five years?
More industries will adopt online identity verification to build trust and aim to automate the manual verification process. AI will play a key role in risk segmentation, user profiling, and building trust based on active interactions with end users.
Trust scoring through an extensive map of signals. More often than not, companies will get that through a single vertically integrated trust provider offering an ecosystem of solutions where identity verification and Biometrics Authentication are the technological pillars.
Behavior-based fraud detection through behavioral analytics. Instead of relying solely on device-based detection methods, businesses can use behavioral analytics to monitor suspicious activity patterns, such as unusual transaction volumes or irregular log-in times.
Be First to Comment