4 top takeaways from the MRC Data Science Summit

Sift recently had the pleasure of sponsoring and speaking at the inaugural MRC Data Science Summit, dedicated to exploring the power of data science and machine learning in fraud prevention and payments. 

The virtual summit was buzzing with conversations about the latest developments in machine learning being leveraged by merchants, the pros and cons of in-house versus outsourced approaches, and some common challenges in data integrity and sharing data while maintaining privacy. 

Speakers, including Bernard McManus from PlayStation, Cicely Robinson from Spotify, and Daniele Micci-Barreca from Google, joined the event with insightful sessions around machine learning and data science for payments and fraud professionals. Among the speakers were Sift’s Kevin Lee, VP of Trust and Safety; Neeraj Gupta, Chief Technology Officer; and Dr. Wei Liu, Engineering Director of Data Science. Here are their top takeaways from MRC.MRC merchant members.

1. Generative AI is changing the game for fraudsters

Generative AI presents both creative potential and heightened risks in the fight against fraud. It floods the internet with disinformation and scams, challenging consumers and businesses to deal with new levels of deception. With tools like ChatGPT, the potential scope of risk is vast, as it can instantly share information and perform tasks at superhuman speed. 

While AI allows scammers to create more convincing content, it also aids in training and improving fraud prevention models. In this ongoing battle, businesses require real-time, comprehensive solutions. Machine learning, augmented by AI, plays a crucial role in identifying evolving fraud patterns that were once undetectable. 

2. It’s time to fight fraud with data and analytics

The 2023 MRC fraud survey reveals a growing need for a technical and analytical approach to fraud management. MRC members’ top challenges center on effectively utilizing data for fraud prevention. To stay ahead of evolving threats, businesses are recognizing the importance of using advanced data analysis tools. This shift emphasizes the importance of a proactive strategy that intelligently leverages data to identify and mitigate fraud risks.

Below, the MRC Fraud survey also highlights two prominent areas that demand attention for bolstering e-commerce fraud management: the enhancement of fraud analytics, and automated detection accuracy. These outcomes underscore the critical role of data-driven insights and precision in addressing the complexities of modern digital fraud. Embracing advanced analytics and refining automated detection systems better equip businesses to fortify their defenses against the ever-evolving landscape of fraudulent activities, ultimately leading to a safer and more secure digital environment for everyone.

3. Sophisticated tools and well-resourced risk teams are a must

Today’s successful risk teams  have to know how to do more than conduct manual reviews, occasionally spot fraud trends, or create rules based on experience. A highly technical approach is crucial, and being skilled in SQL, Python, and Excel has become essential for speed, scalability, and sophistication.

It’s a necessary shift in perspective as trust and safety teams move from reacting to threats to proactively using data analysis and automation to stay ahead. The scale and sophistication of attackers has grown exponentially with the emergence of new technologies, and businesses must set up proactive prevention measures before they quickly become overwhelmed.

4. Machine learning and data science are only two pieces of the puzzle

Automation plays a vital role in rapidly sifting through vast data to detect patterns indicative of fraud. However, it’s crucial to recognize that while machine learning and data science are fundamental elements in a comprehensive fraud prevention strategy, they’re only part of what makes a solution viable. Merchants also need the right people in place to get things done.. 

Well-rounded digital risk teams include data scientists, product managers, machine learning engineers, data engineers, infrastructure specialists, and front/back-end developers, in addition to fraud analysts. This allows them to manage traditional tasks while constructing strong models and workflows, ensuring a proactive approach to fraud prevention.

Moreover, automation plays a pivotal role in empowering business users to apply simple rules without the protracted need for model retraining, which can span weeks or even months. This level of adaptability and agility is essential to stay ahead of the constantly evolving landscape of fraud. In a world where fraudsters continually refine their tactics, the importance of these efforts cannot be overstated, and they are integral to building a robust defense against financial abuse.

If you’re interested in hearing more insights from the Sift team, explore Sift’s upcoming industry events and live sessions.