AI in gambling: Leaders differ on need for specific rules
Industry figures debate whether existing consumer protection law and the EU AI Act are enough, or whether gambling needs bespoke AI rules.
Key takeaways:
- Gustaf Hoffstedt argues gambling laws should regulate outcomes rather than prescribe how AI is used, and says regulators should focus on preventing illegal operators from targeting consumers.
- Catie Di Stefano believes existing arrangements are sufficient for now, saying she has not seen anything alarming about AI in gambling.
- Rasmus Kjaergaard says AI is already used across gambling and argues the industry should not wait for mandates before developing responsible standards.
- Giulio Coraggio says existing law is substantively sufficient but poorly enforced, and warns that operators, suppliers and affiliates may already be exposed under the EU AI Act.
- Craig Cornforth advocates AI supported by trained teams and independent lived-experience assurance, rather than treating AI solely as a risk to regulate.
Special report.- Artificial intelligence is expanding across gambling, from player monitoring and safer gambling tools to marketing, communications and retention models. In the latest edition of Focus Gaming News’ “5 Leaders – 1 Question”, five industry figures debate whether the sector needs an AI-specific regulatory framework or whether existing consumer protection law, enforcement and horizontal rules such as the EU AI Act are sufficient. Their views range from scepticism towards new AI legislation to warnings that gambling companies cannot afford to wait for regulators to act.
The participants are Gustaf Hoffstedt, secretary general of the Swedish Trade Association for Online Gambling; Catie Di Stefano, founder of Campaign Discovery System; Rasmus Kjaergaard, CEO of Mindway AI; Giulio Coraggio, partner at DLA Piper; and Craig Cornforth, senior business development manager at Epic Global Solutions.
Regulate outcomes, not tools
For Gustaf Hoffstedt, well-written gambling legislation should regulate the outcome that needs to be achieved rather than prescribe how it must be achieved.
“Well-written gambling legislation regulates what is to be achieved rather than how it is done. It focuses on a desired outcome and rarely intervenes in how this is achieved. I am therefore sceptical of specific AI legislation.”
Hoffstedt acknowledges that AI can be used with malicious intent to target vulnerable consumers in order to induce them to gamble excessively. However, he argues that “every well-written gambling law already prohibits such actions by gambling companies, whether the actions originate from AI or not.”
He also stresses that AI can be used to improve player protection through increasingly sophisticated and personalised interventions. “AI can of course also be used to develop and strengthen player protection, to help players not to overconsume with increasingly sophisticated and personalised methods. In those cases, too, the legislator should regulate a desired outcome.”
For Hoffstedt, the regulatory priority should be to distinguish between the regulated and unregulated markets. “In the regulated market, AI should primarily be seen as an opportunity to improve consumer protection. In the unregulated market, however, AI can be used in precisely the ways we do not want to see in any gambling market.”
He therefore believes regulators should focus on “preventing illegal operators from targeting consumers, rather than introducing overly prescriptive rules governing the use of AI by licensed operators.”
No immediate need for new rules
Catie Di Stefano believes the existing position is sufficient for the time being. “Right now, I think it’s good as it is. My opinion might change in the future, but today, I haven’t seen anything alarming for AI in gambling,” she says.
Instead, Di Stefano says she has seen AI being used for creativity and problem-solving. While she recognises that the technology has potential drawbacks, she argues that the gambling industry has traditionally adapted quickly to new developments. “AI has its downsides, but as always, the gambling industry is quick to adapt — regulations unfortunately stiffens industry innovation,” she concludes.
The industry cannot wait
For Rasmus Kjaergaard, AI is already embedded in several areas of gambling and the sector should not wait for regulators to force it to act.
He says: “AI is of course used in a variety of ways in our industry, from design and communications to the technology we have at Mindway AI, where AI is combined with neuroscientific insights and expert analysis to provide safer gambling solutions in the way of player detection and monitoring.”
Kjaergaard argues that wider developments in technology and litigation show that questions around algorithmic engagement and user protection are already becoming urgent. “As we speak, we see tech giants like Meta facing landmark litigation over algorithmic engagement and user protection. It is only a matter of time before gambling faces its own bespoke AI safety frameworks,” he says.
For that reason, he says gambling companies should not wait for mandates before developing responsible standards: “The industry cannot wait for mandates to force its hand and react. This isn’t a future risk, it’s today’s reality.”
He then adds: “Leading operators and progressive regulators are already setting the standard: leveraging AI not to drive screen time, but to proactively detect risk, protect players, and embed algorithmic accountability into their core operations.”
Enforcement over new legislation
Giulio Coraggio takes the view that existing law is substantively sufficient, but that it is not being applied effectively. “Existing law is sufficient in substance, but it is not being applied, and a new gambling-specific AI regime would delay the problem rather than solve it,” he says.
Coraggio warns that operators, suppliers and affiliates may already be exposed under the AI Act. “The point most of the gambling industry has missed is that operators, suppliers and affiliates may already be in breach of the AI Act,” he says, and explains, “The AI Act provisions on prohibited AI systems have already been into force for over a year, and since August 2, 2026, the enforcement architecture is complete: EU and national authorities can investigate and impose fines up to €35m or 7 per cent of worldwide turnover.”
Coraggio also points to two areas that could be particularly relevant to gambling: manipulative AI techniques that materially distort behaviour and the exploitation of vulnerabilities linked to economic circumstances. “Two prohibitions land on this sector. Manipulative AI techniques that materially distort players’ behaviour, and the exploitation of their vulnerabilities linked to economic situations,” he says.
Coraggio gives a specific example involving a player identified as being at risk of addiction: “A retention model that spots a player at risk of addiction and releases a bonus at that moment is exposed, as is affiliate targeting built on signals of financial distress.”
He also stresses that responsibility can extend across the technology and supply chain: “Liability runs along the chain, because operators qualify as deployers, CRM vendors as providers, and an operator white-labelling a supplier’s tool becomes the provider itself, with the most stringent AI Act regime applicable to it.”
For Coraggio, the main gap is not a lack of legislation but a lack of implementation and guidance. “Consumer protection law alone would never have reached this, and only after the harm. So the gap is enforcement and guidance from gambling regulators, not new legislation. My advice to clients is to run an AI Act compliance review now,” he says.
AI supported by human judgement
For Craig Cornforth, AI should be treated as an opportunity to improve player protection and operational decision-making, rather than simply as an additional risk requiring regulation.
“AI should be seen as an opportunity for the gambling industry, not simply as a new risk to regulate. The scale and complexity of modern gambling means manual human review alone cannot realistically identify every relevant customer interaction or behavioural signal. Used properly, AI can help operators act earlier, make better decisions and support more customers, while improving consistency and commercial efficiency,” he says.
Cornforth says that Epic Global Solutions is already working with operators to combine AI with lived experience and independent human assessment. “At EPIC, this is already moving beyond theory. We are increasingly working on both sides of the Atlantic with operators to bring lived experience into our EPIC Assurance quality review, with AI playing a role in selecting real-world customer interactions for assessment and our experts validating whether the decisions being made by people and technology work as intended in practice.”
He believes that independent human input is essential to evaluating whether automated systems are working fairly and effectively. “That independent human perspective is critical,” he says, and adds that it gives operators “insight from the very type of customer these systems are ultimately designed to support, while also giving frontline teams greater confidence that their decisions are being assessed fairly and constructively.”
Cornforth’s preferred model combines technology, trained teams and lived-experience assurance. “The strongest model is not AI versus people. It is AI providing scale and speed, supported by trained teams and independent lived experience assurance. That combination can strengthen player protection, improve operational decision-making, and deliver better commercial outcomes,” he says.
