- Joined
- Jan 20, 2004
- Location
- Pictland
AI is turbo-charging iGaming fraud
by Max Drayman, Casinomeister.com“It’s an unfortunate side effect of the artificial intelligence revolution:
a relentless and steep rise in AI scams.”
a relentless and steep rise in AI scams.”
You do not have permission to view link
Log in or register now.
14 April 2026.At this point we are all aware of AI-driven scams: everything from fake bank alerts on your phone, to spoofed phone numbers, fake job offers and dodgy QR codes, to deepfake audio or video of family members.
One significant result is that public-facing services are now faced with a wide-ranging shift in what they receive in their inboxes exactly because of the AI tools now available to the public. A recent survey by Sky News (
You do not have permission to view link
Log in or register now.
) found that in the UK fully 90% of the services offered to the public through
You do not have permission to view link
Log in or register now.
reported that AI-generated submissions to them had had an effect on their ability to deliver their services, with 60% reporting the impact as “significant”. A large part of the issue is sheer volume: those
You do not have permission to view link
Log in or register now.
services reported a 30-70% increase in demand for their service, typically through text-based applications, since the introduction of chatbots in 2022 – the bulk of that coming from Q4 2024 and on.iGaming is now a war zone
In the iGaming industry, we’re seeing AI-driven attacks from all sides: AI-generated spoof sites that perfectly mirror a known brand and run pirated games with hacked RTP; black hole sites that are totally fake casino fronts offering nothing, they’re there to harvest player data from registrations, or they may simply scoop as many deposits as possible before they disappear, only to pop up again under a different name to do the same again. It has been estimated that one of these fraud sites can be created and launched in a few hours thanks to AI-generated code and site copying tools (
You do not have permission to view link
Log in or register now.
).Players are AI-savvy too, running a variety of scams on legitimate casinos, everything from carefully curated multi-accounting on a massive scale to bot-driven account theft at legitimate casinos using purchased or stolen player identities.
Our industry contacts tell us that casino support teams are struggling under the tide of AI-driven fraud. One senior casino representative told us they were losing well more than 100,000€ per month to AI fraud and had no means to restrict or control that… for now.
The view from the PAB desk
Before AI, the fraud scene morphed as time went on, from year to year we could see the shift and adapt accordingly. Now the changes are coming month-on-month if not faster. One week we’ll see a wave of complaints using the same text and tone of presentation. When we learn to filter those out things will typically go quiet for a few days and then the next wave hits: a batch of complaints all from the same country with suspiciously similar problems, or a targeted campaign against an individual casino or casino group, or a single player suddenly has 20 complaints against a range of casinos.These clusters can tell you that it’s all coming from one source, or a group sharing the same procedure, but that doesn’t mean you can simply bin the lot. Are they legit players using the same chatbot, hence the similar wordage, or is it an individual person running a targeted campaign? Is the complaint real or fabricated? Realistic or nonsense? It is often very difficult or impossible to tell until you spend time working the cases.
In late 2022 – the dawn of the AI chatbot – we were handling just under 600 PAB complaints a year and our fraud rate was something like 20%. This year we’ll handle close to 750 cases and we estimate the fraud rate could be as high as 60 or 70%.
Our observation is that today’s fraud cases are often more difficult to prove, take longer to reach a conclusion, and force us to take extra precautions throughout the entire process. We fight to maintain a consistent turn-over time on the cases we handle but it takes a heavy toll on our staff and we’re not always able to deliver a consistent service: sometimes things move slower, and are more frustrating for the honest players who have active cases.
We are adapting, partially by temporarily side-lining obviously meritless, bot-driven cases, but that’s a subtlety that few players appreciate. The unfortunate end result is that we’re seeing more bad ratings for us on TrustPilot and similar ratings sites, both from legit players struggling to express themselves through a chabot who have then been inconvenienced, and from scammers who’ve been caught but can still kick us where it hurts.
Thousands of words that say nothing
When a real player writes a description of their problem they are usually fairly economical in what they say, typically a few hundred words. With AI-generated text, there is nothing limiting the word count: we get flooded with thousands of words that say nothing more, and often much less, than a human’s few hundred words. This is a systemic problem with AI: a
You do not have permission to view link
Log in or register now.
on public services facing the challenges of AI quoted providers as saying “complaints once contained to a single sheet of A4 could now run to 20 pages.” Chatbots are ever ready to generate reams of text.The problem is that nobody is running quality control on chatbot output. When we ask the average person why they said this or that in their complaint submission, they often have no clue what we are talking about. Often, the human never bothered to read the text they had copied from the chatbot, or they have simply forgotten it, or it was never true in the first place (chatbot “hallucinations”) and the story has to be corrected, often repeatedly.
That said, it’s easy to forget that some portion of the population is genuinely helped by access to chatbots. People who do not speak English as their first language or for various reasons are not able to get their ideas across in writing can now use chatbots to help them submit legitimate written complaints.
In order to avoid false accusations, we have to carefully evaluate each submission however bloated. In the age of AI fraud, this becomes a weakness that ends up costing us dearly in time and resources.
Next up: AI agents
In our next installment we’ll look at the latest and most threatening wave of the AI revolution: agents, and the weaponization of AI they represent.Attachments
Last edited:
