AI chatbots recommend illegal casinos and ways to bypass GamStop — investigation
This was reported by DigitalTrends.
A journalistic investigation has revealed that popular chatbots with artificial intelligence sometimes recommend illegal online casinos to users and suggest ways to circumvent gambling restrictions.
During the experiment, journalists tested how different systems respond to requests about betting and online casinos. The study tested tools from major technology companies — OpenAI, Google, Microsoft, Meta, and xAI.
Among the services tested were ChatGPT, Gemini, Copilot, and Grok.
During the tests, some chatbots offered users lists of offshore online casinos that operate without licences and are not regulated in the UK.
A number of the system's responses also contained tips on how to use such platforms.
The researchers paid particular attention to the issue of circumventing responsible gaming mechanisms. The UK has a system called GamStop, which allows people to voluntarily block access to licensed gambling sites.
During testing, some chatbots suggested how to find casinos that are not connected to this system, potentially allowing users to circumvent the restrictions in place.
According to The Guardian, some responses also emphasised features that are attractive to players, such as large bonuses, fast payouts, or the ability to place bets using cryptocurrency.
Such online casinos often operate in offshore jurisdictions, particularly in Curaçao, where the level of regulation is significantly lower.
Regulators note that this may leave users less protected from fraud or problems related to gambling addiction.
After the investigation results were published, the developers announced that they were working on improving their security systems.
OpenAI reported that ChatGPT is designed to refuse requests that could contribute to illegal activity.
Microsoft also noted that its Copilot assistant has several layers of protection designed to prevent dangerous or harmful recommendations.
However, the results of the investigation have intensified the debate about how generative artificial intelligence systems work with sensitive topics — in particular, gambling, mental health, and illegal activities.
In the UK, regulators have already warned that online platforms, including artificial intelligence services, must strengthen their control over the distribution of dangerous or illegal content. These requirements are set out in the Online Safety Act.
