The age of artificial intelligence has well and truly arrived, bringing sweeping change to industry – and gaming is no exception. How are practitioners actively using AI today, and what are the associated risks if we don’t tread carefully?
Artificial Intelligence has been an industry buzzword since long before ChatGPT became a household name, but it is only recently that the technology’s most appropriate use cases have emerged.
And while it undoubtedly promises to be a game changer, AI’s implementation remains measured – a tool to be both embraced and adopted with care.
At a recent Investor Briefing hosted by global supplier Aristocrat Leisure Ltd, the company’s Chief Technology Officer, Bob Serr, detailed how the industry titan is treading the AI path – slowly and judiciously – with the primary objective of boosting productivity.
“AI is not our strategy; it’s an accelerator of our strategy,” Serr explained. Although the technology is seen as an opportunity to grow market share and revenue, he noted that Aristocrat is focused on three key areas: “enhancing creativity, improving velocity to market and advancing insights.”
This, he observed, is no different to any other industry deploying AI for enterprise efficiency and productivity gains.
Perhaps the greatest insight Serr provided during his presentation to Aristocrat’s investors was just how deliberate the company has been in applying AI across the enterprise.
As he explained it, the journey began in late 2022, following the public release of ChatGPT, with experimentation around generative AI – a type of artificial intelligence that follows user prompts to create original content – and how this could be used to create value across the enterprise.
In 2024, Aristocrat expanded its use of AI tools, adopting technologies like Google Gemini, GitHub and Copilot for its software development teams and establishing programs focused on improved productivity and accelerated development.
By 2025, the company “began moving past simple deployment of tools to fully embracing AI,” according to Serr, with processes now reengineered to use autonomous agent-based workflows for increasingly sophisticated tasks. In doing so, Aristocrat was able to achieve “significant and measurable productivity gains [that allowed] our talent to focus on higher value areas,” he said.
On the supplier side at least, workflow improvement has undoubtedly become the single most prevalent use case for AI.
Ian Hughes, Chief Revenue Officer for industry test lab GLI and CEO of GLI Australia, notes that the technology has been particularly beneficial for software developers who now heavily rely on AI tools to write code – historically a highly repetitive task involving extensive manual labor.
“Whether it’s true for the gaming industry I’m not sure, but some studies have found that up to 80% or 90% of a developer’s code base can now be written by AI,” Hughes explains. “That doesn’t mean actually designing games – the design and the architecture and the creative elements are still done by people – but the low-level code that is quite mechanical can now be assigned.
“So, certainly those tools are being utilized by everybody in software development, and if you think about the gaming industry as a subset of software development, there is a huge variety of tools being used by the gaming industry and gaming machine manufacturers.”
The other great benefit suppliers are seeing from AI relates to the testing of ideas. Rather than spending months or even years to build out concepts that are never guaranteed to reach the market – or may already be outdated once they get there – companies are now using AI to build sandpits, do prototyping and undertake basic level development without having to commit substantial resources.
“We’re able to put things in front of customers, put things at trade shows, present the ideas we have and validate those ideas before we assign those precious development hours to a project or to an idea,” offers Konami Australia’s Systems Sales & Operations Manager, Joe Mayer. “That’s been a massive benefit when it comes to innovation, especially on the systems side.”
The emergence of agentic AI – a relatively new technology allowing AI “agents” to be assigned to very specific tasks – has further streamlined such processes. Common uses can range from writing code or fixing bugs to implementing a new software feature or monitoring specific websites for updates (such as regulatory or technical revisions).
As explained by GLI’s Hughes, “Agentic AI is like the general contractor that says, ‘Hey, I’ve got this job that needs doing,’ and has all these agents working for him that are very specific. The idea is that you end up with a better solution because you have each agent doing a very specific task.”
Aristocrat describes agentic AI as part of its “next phase” of AI deployment, with the company looking to leverage this specific type of AI to “redesign increasingly complex workflows at scale,” according to Serr.
On the group’s main focal points, the Aristocrat CTO said he would look to enhance creativity by using AI for ideation, asset creation and general productivity; to improve velocity to market by reducing manual effort across development, testing and quality processes and converting games for additional markets; and to garner greater insights – using AI for analytics and to inform decisions – that support growth and improved player experiences.
“For Aristocrat, AI becomes more valuable when it’s applied to our existing capabilities and strengths and integrated into our current or redesigned workflows to create compounding benefits at scale,” Serr explained. “For example, today, we’re embedding AI into our studio and engineering workflows to increase quantity, quality and speed of content creation. In addition, agentic AI workflows are being deployed to accelerate delivery across channels and markets. And we’re applying AI to proprietary data, providing deeper insights with richer context.”
As an example, Serr highlighted Aristocrat’s social media platform Product Madness, which between 2021 and 2025 increased creative output by 75% – from around 8,000 creative assets to 14,000 creative assets – while simultaneously reducing the number of artists and designers required.
AI is having a similar impact on the gaming floor, where operators are enhancing visibility around their customers’ experience and preferences by using the technology to better track their every move. Depending on the operator’s motivation – or in what part of the region they are doing business – AI can achieve everything from anonymous bonusing and player tracking to clearing regulatory compliance hurdles.
Inside Asian Gaming has written extensively about the growing proliferation of smart gaming tables across Asia-Pacific, now deployed or rolling out at all Macau casinos, at Solaire and Newport World Resorts in Manila, at Singapore’s Marina Bay Sands, at SkyCity in Auckland and in Australia at The Star Sydney and Crown Melbourne.
Put simply, smart tables utilize a combination of radio frequency identification (RFID) and AI to track the movement of chips around a table, attribute them to specific players and ensure all bets are paid or collected accurately. Aside from eliminating dealer error, the technology’s benefits include everything from fraud detection and identification of advantage play to enhanced player evaluation – rating them on their actual betting to give a truer indication of player worth and status.
However, it’s notable that APAC’s two dominant smart table manufacturers, Walker Digital Table Systems (WDTS) and Angel Group, have taken a contrasting approach to AI deployment: WDTS primarily uses AI for the purpose of data analysis by way of AI-driven apps, while Angel’s “hybrid“ system is highlighted by an AI camera system whereby two cameras installed on each side of a gaming table automatically obtain the position, quantity and type of chips in all stacks, capture player actions and ensure that the number of RFID chips in any betting spot matches the total number of chips it sees.
The initial rollout of smart tables was focused on baccarat, but other popular games like blackjack and roulette are now hitting casino floors around the region. Long-term, the goal is for every table to boast full coverage, which would in turn allow operators to not only track every bet placed (chip attribution) but also identify exactly who placed them (bet attribution).
Smart tables are in many ways a nod to the two primary use cases for AI on casino floors today: data analysis and facial recognition.
In Macau, facial recognition has been present for the best part of a decade, although a lack of regulatory clarity has complicated its effective application. The great irony, of course, is that facial recognition first touched the Macau casino industry not via the city’s gaming concessionaires but courtesy of the government, which in 2017 ordered all ATMs – starting with those in and around casinos – to be fitted with a Know Your Customer (KYC) facial recognition system. The measure, aimed at tightening money laundering controls and monitoring capital outflow, specifically targeted holders of China UnionPay cards issued by banks in mainland China.
Two years later, in 2019, Bloomberg reported that facial recognition was being adopted by concessionaires but that Macau’s gaming regulator, the Gaming Inspection and Coordination Bureau (DICJ), had dictated such tools were not to be used to identify high-value players.
A senior executive at one of those concessionaires, who asked not to be named, told IAG that “the regulatory environment in which we work unquestionably constrains the application of AI computer vision.
“The personal data privacy regulations that relate to facial recognition are not clearly defined, leaving the application of the technology somewhat open to interpretation.”
As a result, the executive observed that Macau operators often take a more conservative approach to how the technology is implemented.
Other jurisdictions are becoming far more comfortable with the role of AI and in some places like Australia are even demanding its application as a central compliance tool. For example, facial recognition is used by both Crown Resorts and Star Entertainment Group to identify excluded patrons and prevent them from entering the casino premises. There is a similar push to require Australia’s pubs and clubs – which house a vast majority of the country’s 185,000 poker machines – to adopt facial recognition, and it was recently reported that the NSW state government may in future require facial recognition technology in every gaming room as a means of supporting the statewide exclusion register.
Konami believes the technology’s use cases extend further still, and in recent years the company has been perfecting its SYNK Vision “facial authentication” system to securely identify players when they sit down at a gaming table or gaming machine.
SYNK Vision is the company’s adaptation of computer vision – a type of AI that enables computers to see, understand and interpret images and videos similarly to how humans use their vision.
Described as an “unobtrusive” form of facial recognition, the technology immediately recognizes member individuals who are already logged into the system, while new players are auto-enrolled as a guest account with a rating opened for them.
“It essentially takes traditional facial recognition technology using the AI capabilities of computer vision and algorithms to match live faces against an enrolled face print,” Mayer explains. “That is a true end-to-end AI process.
“With the optimized AI algorithms we have now, we can process all of this at the edge, so it gives us better data security – the face embedding never leaves the camera – and the AI also allows us to do it at a much lower cost.”
At the heart of AI’s rapid expansion across the industry is data, which is not only powering its application but subsequently advancing insights that will dictate the direction industry participants take in the months and years ahead.
In his presentation to Aristocrat investors, Seer noted that understanding how to best leverage that data will ultimately determine the value of those insights and to what extent they can be used to optimize performance.
“This includes a greater understanding of content performance, which informs product strategy and facilitates iteration,” he explained. “It also improves player engagement by enabling personalization on digital platforms, such as through serving the right live ops at the right time or offering personalized recommendations.”
The next level of opportunity that data can unlock, Serr added, is applying AI-powered insights to optimize performance for customers – allowing them to better optimize their slot floors and improve the overall player experience.
But there are risks. As Mayer points out, “For all the operational value you can get from AI, one significant data breach can destroy that value overnight.”
To minimize the threat of any breach, suppliers typically silo their AI use well away from public access software and, if they’re smart, take a similar approach to concept generation versus production.
“Certainly nothing generated by AI ever goes anywhere near production,” Mayer continues on the Konami approach. “We still have our standard development process. We really limit our use of AI to proof of concepts and fleshing out ideas rather than actual end-user development.
“As for data security, we use company-approved internal AI platforms. Nothing goes on public, nothing is general cloud-based use. It’s all very much controlled by our security and IT policies.”
GLI’s Hughes agrees that the biggest risks to data security are putting intellectual property into AI models that are generally accessible, as well as unintentional disclosure by an employee using AI.
“We’ve seen cases of people not understanding the implications using some AI tools and unintentionally putting things out there that aren’t secure, that are visible to open models, and then that learning can be used by competitors to gain insights. You can lose your strategic position because now you’re putting out intellectual property and confidential information.”
As a company with extensive access to the IP of its clients, GLI’s governance framework therefore tightly restricts the AI tools its employees are permitted to use. Those who are authorized must operate within a domain that is hosted on-premises – a process specifically designed to avoid open models that may be accessible to external users.
As Hughes points out, however, risk is not limited to data security, and he cautions against some pitfalls that industry suppliers may face as they increasingly adopt AI to expedite their creative processes.
“The interesting thing is that IP protection doesn’t exist for AI-generated content,” he says. “If you use agentic AI to create a character, and it wasn’t created by a person, there’s no copyright on that.
“The danger is that, because generative AI works by learning patterns from existing data, the genesis of a character design that has been prototyped using AI might unintentionally incorporate a mashup of other peoples’ concepts. It might have used some element of a Flintstones or a Hanna-Barbera character without you even realizing it.
“Those cases haven’t been tested in court, but as the rollout of content gets faster and faster it’s something the intellectual property lawyers will no doubt be watching closely.”
The other worry for Hughes is the death of original content.
“AI works by looking at existing stuff and putting it together in new ways,” he continues. “Even generative AI uses probabilistic methods to determine which words go together, and we haven’t seen AI come up with any new literature. It’s not going to come up with the next Harry Potter novel, it’s just using existing data.
“The less original content that is going out into the world, the more of the same we’re going to see, and I just worry about how this will play out in the gaming industry – particularly creative content. If there are no game designers coming up with new, innovative ideas and you’re just regurgitating, then everything that is released will just be a deviation or a second or third or fourth iteration of something that already exists. That’s going to be sad.”
For all the noise around AI, the industry’s experience suggests its greatest value lies not in replacing the work that people do but in amplifying it. From streamlining software development to refining player analytics and strengthening compliance, AI is proving particularly effective when applied with a clear objective and appropriate safeguards. Be they operators or suppliers, those that benefit the most will likely be the ones who best integrate the technology’s capabilities into their existing strengths.
