AI Fraud and Integrity Detection in Gambling: Match-Fixing, Bonus Abuse, AML
AI now watches betting markets for match-fixing, flags bonus abuse and helps fraud teams investigate accounts. Here is what each named system does, the numbers vendors claim, and what regulators have said about relying on it.
Key takeaways
- 1Match-fixing: Sportradar's UFDS AI monitors odds from 600+ bookmakers. Sportradar says AI flagged 977 suspicious matches in 2023, 73% of all it found, and analysts review every alert before it is reported.
- 2Bonus abuse: EveryMatrix's Bonus Guardian (launched December 2025) flags abuse and can exclude bonuses or hold withdrawals, but no operator is named as using it yet.
- 3Investigations: GeoComply's MCP server, launched September 2026, lets fraud analysts query risk, device and location signals through AI assistants.
- 4Caution: the UK Gambling Commission said in June 2026 that AI anti-money-laundering tools too often fail to deliver what is required, and told operators to prove they work first.
Fraud and integrity work was one of the first places gambling companies put machine learning to use, because the data is huge and the patterns are subtle. Below are the systems named in public documents, grouped by what they protect against. Each has a full profile in our AI products tracker.
| Protects against | System | Who runs it | Evidence |
|---|---|---|---|
| Match-fixing and suspicious betting | UFDS AI | Sportradar, for federations and leagues | Named deployment (UEFA) |
| Bonus abuse | Bonus Guardian | EveryMatrix, for operators | Vendor claim |
| Account and location fraud | GeoComply MCP | GeoComply, for fraud teams | Vendor claim |
| Money laundering and player risk | Neccton mentor | OpenBet, for operators | Named deployment |
Match-fixing: UFDS AI
Sportradar’s UFDS AI is the best-known integrity system. It monitors odds from more than 600 bookmakers, plus account-level staking data and suspicious-bet reports from operators, and Sportradar says its monitoring covers more than a million matches across more than 70 sports. Sportradar offers the platform free to sports federations and leagues.
The figures are Sportradar’s own. In its 2023 integrity report, Sportradar said AI detected 977 suspicious matches, 73% of all the suspicious matches it found that year. Every alert is reviewed by analysts before it is reported to a federation, and our UFDS AI explainer explains why an alert alone is not proof.
On evidence, Sportradar describes the system as “CAS-endorsed”. The record behind that: in Skenderbeu v UEFA (CAS 2016/A/4650), the Court of Arbitration for Sport accepted reports from UEFA’s Betting Fraud Detection System, which Sportradar supplied, as evidence. That was an earlier system, not the current UFDS AI.
Bonus abuse: Bonus Guardian
EveryMatrix launched Bonus Guardian in December 2025 as part of its bonus and engagement tools. It learns from real-time player activity, links to payment and account systems, and can respond by excluding a player from bonuses or holding a withdrawal. EveryMatrix says it reduces manual work and false positives but has published no figures, and no operator is named as using it. In a July 2025 interview, EveryMatrix’s chief technology officer said Bonus Guardian flags issues on a dashboard and a person decides what to do.
Investigations: GeoComply MCP
GeoComply, best known for geolocation checks in US states, launched GeoComply MCP in September 2026. It is a Model Context Protocol server, so fraud and compliance analysts can ask questions in plain language through an AI assistant and get back scored verdicts, linked accounts and the device, network and location signals behind them. GeoComply says bulk lookups take up to 500 usernames or device IDs at once and every action is logged.
Money laundering and transaction risk
Some player-risk platforms also cover anti-money-laundering. OpenBet’s Neccton is sold as a combined responsible-gaming and AML platform that scores risk on live transactions. Its harm-detection side is compared in AI responsible-gambling tools compared.
AI as the threat
Companies also list AI as a fraud risk. Inspired Entertainment’s 2025 annual report names AI used by malicious actors as one of the major risks to its business, because AI-powered bots can imitate human players. The same report says Inspired uses “limited artificial intelligence and automated analytics tools” and does “not use AI to determine game outcomes or to directly influence player results.” Other companies’ statements are on their earnings pages, for example Inspired.
What regulators say
- AI anti-money-laundering tools: The UK Gambling Commission’s director of enforcement, John Pierce, told the GAMLG conference in June 2026 that the Commission is not against operators using AI or algorithms for AML. He said the evidence so far shows they too often fail to deliver what is required, so operators should make sure such tools achieve compliance before relying on them.
- Regulators using AI themselves: In June 2026 the UK advertising regulators began a sweep of gambling content marketing using the ASA’s AI-based Active Ad Monitoring System, announced to operators by the Gambling Commission.
All measures are tracked, with status and dates, in the AI regulation tracker.
Latest on this topic
- EveryMatrix CTO Pablo Jensen: AI already improves Bonus Guardian and game recommendations
- Mindway AI discusses player safety approach and GameScanner application in African markets
- GiG launches PowerPlay platform in Alberta with first deployment of SportX engine
- UFDS AI explained: betting alerts still need human investigation
- UK advertising sweep uses AI to check gambling content
Questions
How does AI detect match-fixing?
Systems such as Sportradar's UFDS AI compare live odds movements across hundreds of bookmakers, together with account-level betting data, against what the market would be expected to do. Unusual patterns trigger alerts, which analysts review before any report goes to a federation or league.
Is AI match-fixing evidence accepted by sports courts?
Reports from an earlier system have been. In CAS 2016/A/4650 (Skenderbeu v UEFA), the Court of Arbitration for Sport accepted reports from UEFA's Betting Fraud Detection System, which Sportradar supplied, as evidence. That predates the current UFDS AI, and Sportradar's own figures for the AI system are self-reported.
What is bonus abuse detection?
Bonus abuse means using promotions in ways the terms forbid, for example multiple accounts or matched betting. Tools such as EveryMatrix's Bonus Guardian learn from real-time player activity to flag likely abuse and apply responses such as excluding a player from bonuses or holding a withdrawal for review.
Does the UK Gambling Commission allow AI for anti-money-laundering?
Yes. In June 2026 its director of enforcement said the Commission is not against operators using AI or algorithms for AML, but that the evidence so far shows they too often fail to deliver what is required. Operators should make sure a tool achieves compliance before relying on it.