Home > iGB@ICE > From evidence to action: inside the ICE Research Institute

From evidence to action: inside the ICE Research Institute

| By Martin Bjoerck | Reading Time: 5 minutes
Two years in, the ICE Research Institute is shifting from funding research to driving its adoption. Director Ewa Bakun explains why, and UNLV International Gaming Institute's Kasra Ghaharian introduces the AI Use Case Registry ahead of ICE Barcelona 2027.

Plenty of gambling research ends in a journal or on a conference stage. Ewa Bakun wants the ICE Research Institute (IRI) to be different.

“While there is a considerable amount of academic research, much of it stops at the point of publication or presentation,” says Bakun, the institute’s director. “We wanted to do something different.”

That ambition is about to be tested. With ICE Barcelona returning to Fira Barcelona Gran Via from 18 to 20 January 2027, the institute is moving out of its first phase – funding studies – and into a harder one: persuading operators, suppliers and regulators to act on what those studies find.

A legacy beyond three days

The institute was born out of ICE’s move from London to Barcelona. Sitting down with the City of Barcelona and Fira Barcelona, the organisers asked what legacy the show could leave beyond its economic contribution, approximately €255 million for the city, and beyond the three days of the event itself.

The answer was an institute focused on the intersection of prevention, policy and technology. Its advantage, Bakun argues, is not academic credentials but reach. Through ICE, iGB Live, iGaming Business and World Gaming Executive, it can put findings in front of stakeholders across the ecosystem and, crucially, work with them to encourage adoption.

“We don’t want to simply finance research and stop at publication,” she says. “Ultimately, we want to generate meaningful, tangible change in how businesses, regulators and other stakeholders operate.”

Prevention, not treatment

The IRI currently funds seven projects, ranging from AI-based early detection of gambling-related vulnerability to adolescent education and media stigma. Reflecting the Barcelona legacy, every project is either led by a Spanish institution or pairs one with international partners, such as the University of Nevada, Las Vegas.

The common thread is prevention. “When you’re looking at treatment, the problem has already occurred,” Bakun says. “We want to investigate areas where the industry can change its practices and therefore influence the impact it has.”

Several projects ask uncomfortable questions. One uses neuroscience to test whether safer gambling messages actually change behaviour; another investigates why relatively few people use harm-minimisation tools.

No fundraising and no gambling industry

“What’s the point of having a harm-minimisation tool or an intervention if it doesn’t make a difference? Or, potentially, if it has the opposite effect to the one intended?” Bakun asks. “Actions should be judged not simply on our willingness or intention to implement them, but on the actual effect they have.”

That only works if the research is trusted. The institute does not fundraise, and no gambling industry money is involved: the ring-fenced fund comes through ICE’s agreement with Fira Barcelona, and academic rigour rests with the universities doing the work. Bakun is now pausing new commissions to focus on completing existing projects, while setting up two independent bodies to oversee future calls and grant selection.

Bringing AI out of the shadows

The first project to put something practical in operators’ hands is the AI Use Case Registry, developed by the UNLV International Gaming Institute’s AI Research Hub with Universidad Rey Juan Carlos.

Its starting point is that AI adoption in gambling is outpacing regulation. Kasra Ghaharian, director of research at the International Gaming Institute, says the inaugural State of AI in Gaming report, published in April 2026, found companies deploying AI everywhere from back-office productivity tools to customer-facing products, but very few with governance processes in place.

“This looks similar to the ‘move fast and break things’ era of social media,” he says, “and it’s something we should really avoid repeating.”

A self-governance tool

The biggest gap, in his view, is visibility. “Without knowledge of where and how AI is being used, it’s going to be very hard to put any governance in place.”

The Registry is a self-governance tool that lets an organisation build an inventory of everywhere it uses AI, alongside documentation of how each use is governed. It is free and open source: companies download it from GitHub, adapt it and run it on their own systems, either as a tool every department uses to log its use cases or under a single AI oversight team.

At its heart is a short screening step that sorts each use case by its potential impact on players and sets documentation requirements to match, across three tiers. “A back-office tool would need only light documentation, while something that has direct interaction with players would require more comprehensive governance requirements,” Ghaharian explains.

To be used right away

That player-impact lens came out of the research. The team reviewed frameworks including the EU AI Act and the UK’s Algorithmic Transparency Reporting Standard. Both, Ghaharian says, rightly scale oversight with risk, but are written for a broad audience and hard to apply in a sector whose product carries addictive potential. Many governance frameworks also stop at high-level principles. “We wanted something practical that companies can implement and actually use right away.”

Regulators, operators, academics and AI specialists then shaped the tool through focus groups and two rounds of testing – which Ghaharian describes as “less a test of a finished product and more a co-design process”. The clearest message was to keep it short: nobody wanted another long compliance obligation.

More than 25 companies using the Registry in a year

Bakun stresses the Registry is not a public benchmarking exercise. “We are not collecting company data or comparing individual businesses against one another.”

Ghaharian’s measure of success is concrete: more than 25 companies using the Registry within a year, and five regulators worldwide recommending it to licensees. Regulators, he says, will increasingly expect licensees to show how they govern AI – and, as with good data privacy practices, which now shapes consumer choice, responsible AI use may become a competitive advantage for companies.

What to look for at ICE 2027

The research will sit at the heart of the new World Gaming Council, a membership proposition developed from the World Regulatory Briefing. Instead of traditional panels, members will receive selected findings first and discuss them at senior level. Four projects have been chosen for detailed consideration, and all seven will be presented in a dedicated session.

More operational sessions in the Sustainable Gambling Zone will focus on AI and neuroscience. Plans for the Registry include a research breakfast introduction, an early-adopter case study and a boardroom discussion on maximising AI’s opportunities while managing its risks. Organisations can access the AI Use Case Registry now and share feedback as early adopters.

“We want visitors to leave not just having heard about the research,” Bakun says, “but with a better understanding of how they can use it.”

The seven projects at a glance:

  1. Stigma and public perception (Episteme Social, Universitat Autònoma de Barcelona): analysed 726 Spanish newspaper articles from 2011 to 2024 and found that framing gambling as a personal moral failing can deepen harm.
  2. GamBrain (Fundación Patología Dual): uses AI on clinical, behavioural and genetic data to detect gambling-related vulnerability early.
  3. Applied neuroscience for prevention (Politecnico di Milano, Universitat Politècnica de València): tests safer gambling messages with EEG, eye-tracking and physiological measures.
  4. Safer gambling tool adoption (University of La Rioja, McGill University, BetBlocker): asks why harm-minimisation tools are underused.
  5. AI Use Case Registry (UNLV AI Research Hub, Universidad Rey Juan Carlos): an open-source framework for documenting and governing AI use.
  6. Prevention versus prohibition (Emooti/University of La Rioja, Proyecto FES): compares prevention with restrictive approaches for adolescents.
  7. Sustainable Gambling Framework, phase 1 (Episteme Social): maps sustainable gambling across jurisdictions, with findings due at ICE Barcelona 2027.

ICE Barcelona runs from 18 to 20 January 2027 at Fira Barcelona Gran Via.

Subscribe to the iGaming newsletter