All five major AI chatbots tested by The Guardian and Investigate Europe could be prompted to recommend unlicensed online casinos operating illegally in the UK, with some offering detailed guidance on how to circumvent player protection controls designed to prevent fraud and addiction.
The investigation, published on 8 March 2026, tested Microsoft Copilot, Grok, Meta AI, OpenAI’s ChatGPT, and Google’s Gemini. Each was asked six questions covering how to find the “best” unlicensed casinos, how to avoid source of wealth checks, and how to access operators not registered with GamStop, the UK’s mandatory self-exclusion scheme for licensed operators.
Every chatbot produced lists of unlicensed casino recommendations. Only two — Copilot and ChatGPT — opened their answers with any health warning. None consistently refused to engage with requests designed to facilitate access to illegal gambling sites.
What the Bots Said
Meta AI, accessible via Facebook, Instagram and WhatsApp, characterised source of wealth checks — mandatory AML controls under UK licensing rules — as a “buzzkill” and a “real pain”. It followed up with tips on how to avoid them, recommended specific unlicensed sites based on their bonus offers and cryptocurrency payment options, and described one operator’s gameplay as offering “generous rewards”. No UK-licensed gambling operator is permitted to accept cryptocurrency payments.
Grok advised users that paying with cryptocurrency would allow “funds to go directly to/from your wallet without linking to bank accounts or personal details that could prompt verification”. Gemini told users that offshore casinos offered “significantly larger” bonuses than their licensed equivalents, and provided a step-by-step guide to accessing unlicensed sites — though it subsequently refused when the question was repeated in a second test.
ChatGPT produced a side-by-side comparison of unlicensed operators, ranking them on bonuses, game libraries, cryptocurrency versus card payment options, and payout speeds. OpenAI disputed the characterisation, stating the bot was “trained to refuse requests that facilitate behaviour” and that it had done so by “providing factual information and lawful alternatives”.
Microsoft Copilot described unlicensed casinos as “reputable” or “trusted” in its output. A company spokesperson cited “multiple layers of protection, including automated safety systems, real-time prompt detection, and human review”, adding that these safeguards were continually evaluated.
Google said Gemini was “designed to provide helpful information in response to user queries and highlight potential risks where applicable” and that it was “constantly refining safeguards” to ensure complex topics were handled appropriately. Meta and X did not respond to requests for comment.
Regulatory and Compliance Implications
Unlicensed operators typically shelter under licences from low-scrutiny jurisdictions such as Curacao to maintain a nominal regulatory veneer, while targeting players in markets where they are not authorised to operate. The UK Gambling Commission’s final report on the illegal online gambling market documented the scale of unlicensed activity targeting British consumers and the difficulty of enforcement without cooperation from payment processors, ISPs, and platform operators.
The Gambling Commission confirmed to The Guardian it “takes this issue very seriously” and is participating in a government taskforce examining how tech companies can be held responsible for harmful or exploitative content circulating on their platforms. The UK government pointed to the Online Safety Act as the legislative mechanism requiring platforms to remove illegal content, and said it would “not hesitate to go further” if evidence of ongoing harm warranted it.
The cryptocurrency angle in several chatbot responses adds a specific regulatory dimension. The Gambling Commission has been accelerating its cryptocurrency regulatory timeline as adoption rises among younger demographics — a user group the Commission has identified as disproportionately exposed to unlicensed gambling products. AI chatbots actively steering users toward crypto-payment unlicensed sites compounds the enforcement challenge.
Henrietta Bowden-Jones, the UK’s national clinical adviser on gambling harms, said no chatbot should be allowed to promote unlicensed casinos or undermine GamStop, describing the scheme as a free protection service that allows people to block themselves from gambling sites.
GamStop and Self-Exclusion Exposure
Several chatbots specifically answered queries framed around bypassing GamStop — the very protection mechanism mandated for all UKGC licensees. This is a significant compliance signal for B2B operators whose platforms may be exposed to AI-assisted traffic flows. If AI chatbots serve as a discovery layer for users actively seeking to circumvent self-exclusion, it raises questions about whether current Know Your Customer and AML controls are adequate to identify such users at onboarding.
The Gambling Commission has issued £18m in penalties across 2025 for compliance failures, with player protection failures among the most cited breach categories. The emergence of AI as a referral channel to unlicensed gambling is likely to feature prominently in the Commission’s next review of its enforcement priorities.
Chloe Long, whose brother Ollie died by suicide in 2024 following struggles with gambling addiction linked in part to illegal casino use, called for stronger regulation and accountability for platforms that direct people toward illicit sites. An inquest into his death found unlicensed casinos were part of the circumstances leading to his death.
What Comes Next
The investigation does not name specific unlicensed operators, but its findings present a material compliance risk for the AI platforms involved and a policy problem with no straightforward technical solution. AI models trained on internet-scale data will naturally encode knowledge of unlicensed operators, and output filtering at inference time has clearly not been sufficient to prevent harmful recommendations from appearing in user-facing responses.
For the broader iGaming industry, the concern is twofold: unlicensed operators gain a free discovery mechanism via mainstream AI tools, and licensed operators face a competitive disadvantage as chatbots emphasise the larger bonuses and lighter KYC requirements of illegal sites. The UK government has indicated legislative review is possible if platform compliance under the Online Safety Act proves insufficient — a timeline that will depend heavily on how rapidly tech companies move to address the gaps the investigation has exposed.
Source: The Guardian / Investigate Europe
