Turkey Uses AI to Trace Illegal Betting Networks Across 39 Provinces

Turkish authorities have turned an Istanbul police-built AI system, AVCI, into the main tool for tracking alleged illegal betting networks running on Telegram and other digital platforms. For PSPs and payment teams, the point is not the software itself but the trail it leaves: accounts, messages, devices, bank accounts, crypto accounts, and sites that can be tied to a betting ring.

1. Police said the nationwide operation led to raids in 39 provinces on November 3. Investigators initially targeted 194 suspects and arrested 169 people, while court orders later put 121 suspects behind bars pending trial. Another 45 were released under judicial supervision and three without restrictions, according to Turkish Minute.
2. The core of the investigation was AVCI, an open-source intelligence analysis system developed by Istanbul police. The software uses natural language processing to spot coded language, slang, and emoji combinations in online chats, and it can also review photos and videos, link accounts under different names, and map relationships between suspected members of a network.
3. Authorities said AVCI helped identify around 306,000 potentially incriminating messages in 47 Telegram groups in the latest case. One suspect was described as a 21-year-old software developer and a central player in the network, allegedly using a Telegram identity that referenced British lottery winner Michael Carroll.
4. This was not a one-off. In a previous operation in May, police arrested 108 people in 35 provinces after reviewing around 600,000 messages with AVCI. That case resulted in the blocking of 5,151 websites, including 5,000 illegal betting platforms, 111 promotional websites, and 40 sites said to be processing payments.
5. A separate gendarmerie-led probe in Aydın found 40 suspects whose bank accounts recorded TRY 3.484 billion ($102 million) in transactions. On November 3, authorities detained 37 people in raids across 14 provinces and imposed precautionary measures on 467 bank accounts, four cryptocurrency accounts, and four vehicles; phones, computers, storage devices, and SIM cards were also seized.

For high-risk operators, the practical signal is clear: Turkish enforcement is not just looking at betting sites, but at the surrounding payment and communications infrastructure as evidence. If a network can be reconstructed from Telegram chats, bank flows, and crypto touchpoints, the payment side stops being a back-office detail and becomes part of the case file.