Fraud Detection in Online Sports

Business Problem

A 2019 survey conducted by a leading gambling fraud software provider found that 2 in every 5 respondents have closed an online gambling account at least once due to fraud and/or privacy concerns and two-thirds of respondents indicated that they would switch providers based on the quality of their security protocols. Trust is a core component of any transactional platform, but the novelty and scrutiny around legal gambling may make the trust component particularly vital. It also has mass appeal and exciting market growth potential. As of Summer 2019, online gambling accounted for 11% of internet traffic and continues to grow – the market size is projected to reach $1 trillion by 2023, which would represent around 1% of projected global GDP. With extremely large amounts of cash and digital assets riding on the integrity of the security systems of leading online gambling platforms, developing highly intelligent processes for detecting and addressing fraud is an essential risk management measure. Beyond the risks players face concerning account takeovers and credit card fraud, the casino itself bears much of the remaining fraud risk; tactics such as bonus abuse (collecting reward money for referring fraudulent/fake accounts), chip dumping / gnoming (playing from multiple accounts to drive up jackpots and bonuses), and algorithmic betting / bots all expose the casino to additional risk. As hackers plot increasingly sophisticated cyber attacks and players find new ways to cheat the system, having holistically strong infrastructure for fraud detection will be crucial to the survival and success of new and incumbent market competitors.

Intelligent Solution

Assessing fraud risk is a proven AI use case across industries, and the growing legal gambling industry can apply this to its transaction logs (deposits, withdrawals, etc.), betting patterns, and other relevant metadata to predict anomalous behavior at a highly granular level through supervised and unsupervised machine learning methods. AI is a great tool for solving problems with the level of complexity and scale and helps casinos keep up with the latest adaptations of fraud techniques. By assessing characteristics of the player’s background, previous betting behavior, IP Logs, etc. AI can synthesize historical transaction logs and auxiliary data sources to uncover leading indicators of fraudulent activity that would otherwise go undetected.