Artificial intelligence can do a lot. In the hands of an online platform, it can predict when a user is most vulnerable and use this knowledge in two fundamentally different ways. The first is to send a personalized offer precisely when a person is least likely to refuse it. The second is to offer a break precisely when it's most needed.
Technically, these are the same algorithm. Ethically, they are opposite decisions.
Andrey Dobrovolskiy, beneficiary of Neuralink (Cosmobet brand), works with precisely this choice. Let's look at what responsible use of AI in iGaming looks like from the inside, and why this issue has long ceased to be exclusively industry-specific.
What the algorithm knows about the user
Modern analytics systems in online gambling work with data that most users are unaware of. This is a behavioral profile, the accuracy of which is comparable to what a close friend knows about a person, and in some respects, even superior.
The algorithm records and analyzes:
- Bet frequency and size, including the dynamics of changes after wins and losses;
- Session duration;
- Decision-making speed;
- Time of day and regularity of logins;
- Response to notifications and offers.
Most digital products (not just gambling) use this data to optimize engagement. The algorithm learns to identify the moment when a user is most ready for a target action and acts accordingly. This increases conversion. This creates an environment in which the user systematically loses information asymmetry to the platform.
Cosmobet formulates an alternative approach through specific principles:
- Algorithm transparency;
- Influence limitation;
- Control over offers;
- Risk response.
"Analytics and artificial intelligence algorithms are part of a responsible approach to the user. The implementation of AI solutions should adapt the product to the needs of different groups of players and promptly respond to risky behavior patterns."
Preventive Protection
Traditional user protection logic in gambling is reactive. Verification upon registration, the ability to self-exclude upon request, and support upon request. All of this is necessary. But it only comes into play after the situation has become critical.
The algorithmic approach opens up a fundamentally different, preventive, opportunity. The system can detect changes in behavioral patterns long before the user even realizes what's happening.
What this looks like in practice:
- Real-time activity monitoring;
- Personalized warnings;
- Assessment of a shift to impulsive gambling;
- Automatic pause after intense sessions.
"Platforms can go beyond simply complying with regulatory requirements and create a responsible environment where player safety becomes part of the overall product approach, not a separate feature in the settings section."
Andrey Dobrovolskiy
The fundamental difference between this approach and standard compliance is that the protection tools are integrated into the product logic, not added on top of it. This requires operational investments and reduces short-term engagement rates.
Algorithm Ethics
The debate about the ethical use of AI in digital products is actively underway in the context of social media, fintech, and advertising technology. In iGaming, it is noticeably less public, despite the fact that the stakes are literally higher.
The problem is that the industry has yet to develop uniform standards for their application in the context of responsible gaming. Regulators require certain protective tools. But they do not yet regulate how AI is used to influence user behavior in the interim.
This is the real ethical dilemma. Dobrovolsky's position here is clear:
"Even if a platform has legal grounds for implementing behavioral models, this does not relieve it of responsibility for the consequences. The key challenge is balancing when technologies promote transparency and when they covertly encourage repeat actions."
The question is not whether AI will play an increasingly important role in shaping the user experience. The question is who sets the boundaries of its application and how.
While regulators are still formulating requirements for this area, operators are faced with a choice, wait for the boundaries to be set externally or define them themselves. Cosmobet and Andrey Dobrovolskiy's approach represent the latter option. They argue that this choice is possible and that it is best to make it before it becomes mandatory.
