SPORTS BETTING
Brazil experienced deepfake fraud at rates five times higher than the US and ten times higher than Germany, which means a one-size-fits-all compliance model simply doesn’t make sense, and controls need to adapt to local risk realities.
GIO: How should iGaming platforms decide what to automate versus what still requires human oversight when it comes to fraud detection and compliance workflows? KG: Automation should be applied wherever decisions are repeatable, rules-based and high-volume, essentially in areas where speed, scale and consistency matter more than subjective judgement. That includes much of the operational layer such as identity verification, document authenticity, data extraction and cross-checks, liveness and biometric matching, sanctions and PEP screening, behavioural pattern detection, and the generation of transaction monitoring alerts. In these cases, machines are more reliable at scale, and obviously faster than humans.
Human oversight should remain at the point of consequence. Final risk assessments, customer classification, borderline or complex cases, and any decision that could result in account suspension, closure or regulatory reporting need human judgement, context and accountability. The goal is ultimately to surface risk across KYC, transactions and behaviour, so analysts can make better decisions, faster, and with confidence.
GIO: Affiliates play a critical role in player acquisition, yet they’re often left out of fraud conversations. What risks and responsibilities do affiliates need to be more aware of? KG: Fraud goes both ways between operators and affiliates without doubt, but traffic is definitely heavier in one direction than the other.
While ‘affiliate fraud’ typically describes the type of traffic operators consume, with the affiliates often being associated with being the baddies, affiliates themselves need to be careful too.
Fraud against affiliates normally materialises in the form of shaving, where valid players are tracked, but deposits or LTV is partially or selectively excluded. I’ve also known cases where affiliate terms and conditions have been changed retroactively after traffic has been sent, and even though it’s rare, instances where players have been de-tagged as well. When we consider all this, as well as the layer of affiliate-on-affiliate fraud, an unfortunate takeaway is that despite these threats, affiliates have the least visibility, protection, or regulatory backing within the industry.
GIO: With regulations varying dramatically across jurisdictions, how can platforms design fraud and compliance systems that scale globally without becoming overly complex? KG: It’s not just regulation which varies dramatically across jurisdictions, but also fraud. In the past year, 83% of industry professionals reported a rise in fraud, with one third estimating it costing their business 10-20% of total annual revenue. The regulation and anti-fraud side of the industry has become so nuanced, that it makes more sense, and becomes more cost effective, to work with a partner who specialises in this area (there’s one in particular that comes to mind!) than it does to manage it all in-house.
Scalability and adaptability need to be at the forefront of an operator’s strategy. You don’t need to just be able to adapt on the fly when you enter new markets but be able to pivot when markets you’re already in change. Whether that’s previously unregulated markets which subsequently regulate, or whether it’s somewhere like Buenos Aires, notorious for making regulatory changes at short notice. You need a dashboard which allows you to toggle compliance features on and off by region, create and adapt risk scores, build workflows for onboarding, as well as having the power behind it to detect constantly evolving AI threats and supporting features, such as for example detection of consistent backgrounds across selfies, which link into device intelligence, and related user groups.
GIO: AI and machine learning are widely discussed in fraud prevention. Where do you see genuine value today, and where is the industry still overhyping the technology? KG: The value right now is in the ability to spot patterns humans simply can’t at scale, behavioural anomalies, device inconsistencies, network signals, and subtle correlations across sessions, accounts and geographies. That’s where machine learning genuinely outperforms rules alone. Where the industry still overhypes the technology is in the idea of fully autonomous decision-making. AI isn’t a silver bullet, and without strong data quality, human oversight and clear accountability, it can just as easily automate bad decisions. The real value comes from augmentation as opposed to replacement. This is represented in using AI to surface risk faster and more accurately, while keeping humans firmly in the loop for judgement and governance.
GIO: For startups and emerging iGaming platforms, what are the biggest mistakes you see when fraud prevention is treated as a ‘later problem’?
KG: I think the first big mistake is treating fraud as a ‘later problem’ in itself. Startups can be the most vulnerable victims of this. They often need to stand-out versus their competitors, and that often entails splashing their front pages and ads with large, attractive bonus offers, which in turn paints a target on their backs for fraudsters who immediately target them for abuse. Without the correct tools they’re also an easier prey for money laundering, whether in the form of layering, smurfing, muling, opposite betting, arbitrage betting, or any of the other many forms of doing this, they simply aren’t as sharp when it comes to detecting this as the large corporations, and therefore more vulnerable, and the fraudsters know this.
It’s especially relevant when we consider that the top five biggest fraud threats reported by iGaming professionals according to Sumsub’s 2025 ‘State of Identity Verification in the iGaming Industry’, are bonus abuse, money laundering and identity fraud. If you’re wondering, the other two in fourth and fifth places are payment fraud and account takeovers.
GIO: Looking forward over the next two to three years, what changes do you expect in how regulators, operators, and technology providers approach fraud prevention and compliance in iGaming? KG: I expect AI is going to play a bigger role, this will both create and solve a lot of new challenges. I expect we’re going to see a rise in the types of games which have led to some of the next-gen operators growing so fast in popularity. Games such as dice, plinko, hi-lo, as well as gameshows, PvP games, and a refurb of the traditional casino ecosystem. This in itself comes with its own fresh set of risks, including new forms of opposite betting, collusion, advanced money laundering opportunities, XP farming, and much more…
As the iGaming landscape changes, it’s normal to expect fraud to evolve with it. But when you add AI into the equation, you’re supercharging it on both sides. Regulators will also have to regulate for AI. We have tools in place to protect users from, for example, problem gambling, from being unfairly exploited, from betting too quickly, etc.
But will these frameworks evolve to address the new tension whereby AI is capable of optimising engagement, retention and spend at a level that sits uncomfortably close to the very RG principles regulators have spent years enforcing?
GIO FEBRUARY 2026 33
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38
