It’s not easy to draw a full audience at 9 a.m. on the first morning of G2E, but the “Inclusive AI: Development, Progress, and Challenges a Year Later” panel did just that. A year after their original “Women in AI” session, several of the same speakers returned to unpack how far artificial intelligence has come and where gaming still needs to evolve.
The discussion was led by Anastasia (Staci) Baran, Director and COO of nQube Data Science, and she was joined by Dr. Brett Abarbanel, Executive Director of the UNLV International Gaming Institute; Jamie Shea, EVP of RMG Strategy and Marketing at SimWin; and Tamara Hansen, Director of Business Development, FinTech for NRT Technologies, who stepped in for Christina Thakor-Rankin.
Baran opened by noting that AI’s role in gaming is changing quickly and the focus has shifted from potential to practice. Over the past year, operators, suppliers, and data firms have moved beyond experimentation to serious conversations about bias, transparency, and commercial value.
Inclusion as a business imperative
Baran reminded attendees that inclusion isn’t just a moral principle, but a performance strategy. Companies that reflect the diversity of their customers, she said, make better decisions and earn stronger engagement.
She cited examples outside gaming to underline the point. When Pinterest introduced inclusive filters to its platform, engagement rose 66 percent. A study by Accenture found that companies with deliberate accessibility programs consistently outperform those without them. “To increase revenue, we have to understand who everybody is and properly account for diversity in our data.”
Profitability and representation
Abarbanel said inclusion and profitability in gaming are often linked by necessity, but implementation remains inconsistent.
“In many ways, because so much is driven by profitability, inclusivity is sort of innately included,” she said. “If you’re thinking about who all your end users are, then you should automatically be including all of these ideas. But I see the same companies do incredible things for inclusivity and then launch products that completely miss it.”
She added that the industry often rewards volume instead of value. “Don’t use frequency as a proxy for value. The ones performing best are thinking about who the end user is.”
Expanding the audience
Shea said sportsbooks are facing that challenge directly. “If you only talk to one group, then you’ll only get one group,” she said. “By its nature, sports betting has that reputation of bro culture and we’re all trying to get away from that. But at the same time, we don’t want to whitewash it either. We have to find the mix that works for everyone.”
She described meeting a woman building a fantasy sports platform designed for women and beginners. “It’s not that we don’t understand stats or how things work,” Shea said. “It’s that we want something that resonates.”
Hansen added that expansion should never come at the expense of authenticity. “There’s always the desire to reach as many people as possible, while staying true to your purpose,” she said. “You can’t be everything to everyone, but you can continue to expand and do more while staying authentic.”
Bias begins at the data source
The group agreed that bias starts long before a model is deployed. With casinos sitting on some of the most complete customer datasets in entertainment, gaming has both an advantage and a responsibility.
“Bias can creep in before you even start developing,” Baran said. “Before you build a product, you have to understand where the data is flowing from and how systems connect. You can’t begin to address bias until you know the landscape.”
Abarbanel added that even well-intentioned analysts can create problems. “You can introduce bias by trying not to,” she said. “Analysts who don’t understand what the data means are still shaping it.”
Shea pointed out that the problem can lie in the data itself. “If you talk about sports betting, the bias is there, because that’s who the traditional bettor has been,” she said. “But the customer has evolved. We can’t keep asking the same questions we asked in 2018.”
AI as a workforce tool
While risks dominate headlines, the panel also discussed how AI is quietly improving internal operations. Shea said AI-based simulations are already being used to train employees to handle customer-service interactions. “I don’t want to role-play an irate customer,” she said. “AI does it better. It saves time, takes out bias from the trainer, and lets teams practice handling situations they might never see otherwise.”
Abarbanel highlighted new tools such as Google’s Genie 3, which creates adaptive training environments. “It’s a world-building tool with memory,” she explained. “Once you do something in it, that thing sticks. It can recreate real-world continuity without the awkwardness of live role-play.”
Deepfakes, data, and responsibility
Not all uses of AI are positive. Hansen pointed to the rise of deepfakes. “You can submit images of your face and have it do literally anything,” she said. “Those videos were supposed to stay contained — and they haven’t.”
Shea shared a personal story about misinformation. “My mother was recently diagnosed with Alzheimer’s. Friends sent videos of ‘doctors’ claiming they’d found a cure. She believed them. It wasn’t real. It was AI.”
Abarbanel said similar risks exist in mental-health applications. “There have been cases where AI told users how to carry out suicide plans,” she said. “Illinois banned AI for therapy except for recordkeeping. That’s how serious this is. But with proper design, it could also provide real help.”
A realistic path forward
Baran closed the session by reminding the audience that skepticism toward new technology is nothing new in gaming. “When slot machines went from reels to screens, people said, ‘I can’t trust this.’ Now no one thinks twice.”
Shea ended on a lighter note. “My husband is Swedish and very direct in emails,” she said. “He runs them through AI to make them sound polite. It works.”
The takeaway: AI’s evolution is accelerating, but its progress will depend on who’s included in the data — and who’s paying attention to the people behind it.
