New Frontiers in AI for Fraud Prevention

Enjoy the article. Like, share and subscribe top right corner. Thank you!

New Frontiers in AI for Fraud Prevention: Insights from Money20/20's Power Panel

In 2023, global fraud reached a staggering $485 billion, a figure that underscores just how critical it is to tackle fraud more effectively. This massive sum represents only the visible damage; the true cost runs much deeper, impacting businesses, consumers, and institutions worldwide. As the arms race between fraudsters and defenders continues, AI is emerging as the most powerful tool yet in turning the tide.

At Money20/20, the panel titled New Frontiers in AI for Fraud Prevention, sponsored by NVIDIA in partnership with Oracle, brought together three influential leaders to discuss how the latest advances in AI are changing the game in fraud prevention. The session, led by Naré Vardanyan , CEO & Co-Founder of Ntropy , featured Barkha Saxena , Chief Data, AI & Analytics Officer at Chime , and @Kimberly Sutherland, VP of Fraud & Identity Strategy at LexisNexis Risk Solutions.

The Women Leading the AI Charge in Fraud Prevention

Naré Vardanyan set the tone for the session by sharing her unique background in financial inclusion and machine learning. From her early career at the United Nations to her current leadership at Ntropy, she has focused on parsing vast amounts of financial data to fight fraud. Under her guidance, Ntropy has scaled from processing 20,000 transactions to hundreds of millions weekly, parsing billions in USD value. This perspective framed a robust conversation on the future of fraud prevention.

Barkha Saxena: Pioneering Data Infrastructure at Chime

Barkha Saxena’s experience spans from building a data-first culture at Poshmark to leading data initiatives globally at McKinsey. Now at Chime, she’s laser-focused on leveraging data and AI to protect the financial security of Chime’s members, many of whom are financially vulnerable. Notably, Chime's investment in AI has paid off, achieving a 50% year-over-year reduction in account takeovers.

“Fraud prevention has always required cutting-edge models and technology,” Saxena noted. The arrival of generative AI has further transformed the space, allowing teams like hers to process unstructured data at scale, generate real-time insights, and boost the efficiency of fraud analysts through workflow automation.

Kimberly Sutherland: Championing Adaptive Models at LexisNexis

With almost two decades at LexisNexis Risk Solutions, Kimberly Sutherland is an expert in fraud analytics and identity verification. She emphasized that while advanced models have been around for a while, the current focus is on making these models more adaptive. “Fraudsters are using the same technology we are. It’s a constant challenge to stay ahead,” she pointed out.

Sutherland highlighted the growing importance of collaborative intelligence across the industry. LexisNexis is seeing increased willingness among financial institutions to share fraud data, enhancing collective defenses. This collaborative approach helps create a more comprehensive web of signals to detect fraudulent activity more effectively.

How Generative AI is Revolutionizing Fraud Detection

The panelists agreed that generative AI brings unique capabilities to the table. Its ability to generate realistic synthetic data for training models allows for simulation of fraud scenarios that teams might not have real-world data on. Additionally, generative models offer unparalleled opportunities to integrate multiple data sources, automate insights, and make fraud detection not just more efficient but also more proactive.

Barkha Saxena explained that at Chime, the real value lies in blending proprietary data with real-time processing and rapid training of in-house models. The need for comprehensive data infrastructure is paramount, ensuring data flows seamlessly through the pipeline to enable rapid decision-making.

Kimberly Sutherland echoed these sentiments, adding that using synthetic data, particularly around synthetic identities, has been a powerful tool for preemptive fraud detection. “We can better detect fraud by synthesizing data that mirrors real-world, complex scenarios,” she said.

The Challenges: Latency, Accuracy, and AI Hallucinations

Despite the advantages, challenges remain. Saxena pointed out that while models are rapidly improving, they are not perfect and should not be trusted blindly, especially when customer-facing. Human oversight is still needed, particularly when it comes to high-stakes scenarios in financial services. Real-time latency, accuracy of insights, and the ongoing issue of AI hallucinations remain hurdles.

Kimberly Sutherland emphasized the potential pitfalls of AI-generated errors, noting that even small mistakes can have outsized impacts when interacting directly with consumers. “You can’t present wrong information in real-time. It undermines trust and can create new vulnerabilities,” she stated.

Building for the Future: Collaborative AI and Shared Intelligence

The panel concluded with optimism for the future, underlining that the key to effective fraud prevention lies in leveraging AI to augment human decision-making and investing in data infrastructure. The consensus was that as fraudsters become more sophisticated, the industry must advance in step, using AI not just as a tool, but as an integral part of strategy.

“There’s no finish line in this race,” Saxena concluded. “But with the right data, infrastructure, and collaborative efforts, we can be confident in our ability to stay ahead.”

Money20/20’s New Frontiers in AI for Fraud Prevention was more than just an exploration of current capabilities; it was a call to action for the industry. As AI continues to advance, so too will the methods of those seeking to exploit it. The only way forward is to innovate faster, collaborate more deeply, and remain vigilant.

Looking to launch or invest in Fintech ?

If you're an entrepreneur or investor focused on fintech startups, now is the perfect time to reach out and explore investments and expansion opportunities.

Whether you're launching a fintech startup or looking to invest in the AI and fintech sectors, reach out, happy to connect and explore.

Like, share and subscribe top right corner. Thank you!

=============================================================

SUBSCRIBE AND NOT MISS ANY NEW POST !!

If you like this article please share and subscribe (hit button at very top, right-hand corner). Give a shout out to Fintech Next Newsletter, Sheela Ursal, follow or connect with her on LinkedIn or Twitter. Checkout the Fintech Next website and LinkedIn page. Join us at our next EVENT, sign up.

Let's keep building and empowering the Fintech Community!

=============================================================

This post is a reflection of my personal independent opinions and does not reflect views, strategies, and road-maps of my employers, colleagues, investments, companies consulted or mentioned in the article.

=============================================================