The casino industry online and on land has been using artificial intelligence (AI) for quite some time now. However, with the advent of Generative AI across several models, the term with broad meaning can now be used to generate images directly from text prompts.
With competition for your gambling dollar strong, and an ever-increasing number of people around the globe partaking in risk-based entertainment along with a variety of other reasons we’ll look at as this report progresses, it comes as no surprise that AI-generated images have found their way onto slot reels, albeit only on virtual online slot machines at this point.
April 2023 saw the release of at least three online slot titles with imagery created by or with the help of artificial intelligence. The three pioneering companies and titles we’ll cover here are NetGaming (Wild Buffalo Bonanza), Endorphina (Joker Ra), and BGaming (Alien Fruits). And while there could be more (disclosed or undisclosed) we’ll focus on these games as they all came out about the same time and have been heralded in the press – or at least press releases – as innovations and game changers.
Click on the following links to see release notes and details of the following slot titles developed with AI-assisted imaging:
Wild Buffalo Bonanza (released by NetGaming on April 13, 2023, to Whitehat Gaming operators and globally on 4-27) See here.
Joker Ra (released by Endorphina 4-27) See here.
Alien Fruits (released by BGaming on April 13, 2023) See here.
A Little Background on AI in the Gambling Industry
One of the most obvious ways AI has been used in the gambling world is for security Facial recognition coupled with AI has been used to identify card counters, self-excluded gamblers, banned cheats, and organized criminals from land-based operations. On the negative side, it has been used in whale playgrounds like Macau to identify high rollers in order to target them for casino profits through various tradecraft.
On the online end of things, AI can analyze data on player patterns and habits and when an aberration occurs even real-time play can be examined more closely by the machine or by operations management. If it appears the purported gambler is not behaving like the known player, steps can be taken to prevent fraud including theft from the player or casino such as money laundering.
AI is also used more extensively all the time in jurisdictions with robust player protection frameworks to identify problem gamblers and either intervene automatically or by alerting staff who know what to look for, what type of questions to ask, and which actions to take to protect vulnerable persons.
Artificial intelligence has been used by companies big and small in all online markets analyzed. There isn’t much public domain information available but reading between the lines can be productive when the story seems to be the same or similar if what is seen in the gaps is the same across multiple platforms, markets, and jurisdictions.
In a nutshell, the most obvious use of AI in online casinos is in collecting data on players’ behaviors and then shaping the online experience to give gamblers the most satisfying experience possible without affecting the odds or any other element of a fair game.
Understanding what type of slots most appeal to a certain player and then making similar titles with some of the same or similar maths or mechanics will allow the operator to further define what the player is after. Was it simply the buffalo theme or was it the risk and reward factor? Was it the fact there are two bonus rounds or the possibility to buy one of the bonuses?
Just a Few of the Things Affected by AI
These are but a few of the elements that can be sussed out to deliver a better player experience. A better experience means more intense engagement and deeper engagement means longer playtime. Since the House always wins in the end (or at least can maintain an edge until time guarantees it if players don’t quit while they are ahead), the longer a player plays the more they will lose, on average.
The type of bonus you use most often, the sort of wagering requirements you normally accept, whether or not you are more apt to take a once-in-a-lifetime huge cashout type offer or are simply happy to greatly extend playtime and are willing to accept a huge percentage match in exchange for liberal terms and a maximum cap on how much you can win – these can all be handled by AI. In fact, they are and have been for nearly a decade in some instances – but again – it’s not really something that operators or marketers want their competitors to know or even their players – that could take the fun out of it for some.
While many things are accomplished by algorithms or outright AI in the background, none can change the end result of games like slots. Well, sure they could, but not in a way that matters. Regulators such as the BGC, (UKGC) not only require a game to have a stated or “theoretical” RTP (return to player percentage) they also require the actual RTP to fall in line with that given number.
How long (how many iterations or spins) it takes to reach it, how long it stays at that point, and how often and far it deviates over the course of time are simply matters of volatility rather than payback percentages and that is taken into account as well. If it takes a low variance title 400,000 spins for the actual RTP to arrive at the theoretical RTP number – and stays within the standard deviation (say 5 points above, 5 points below, or anywhere in between) for 95 out of 100 simulations (5% confidence interval) then it would appear not to matter to the regulator just how that occurred as long as the maths are solid.
By the same measure, players shouldn’t care how it occurs as long as it occurs – and – each spin has the same chance to produce any given result.
Thankfully, there aren’t any games that mess around with the maths that are written in stone. There was one product on the market several years ago that was really a scratch card but looked like and behaved like a slot machine. The problem with it was that as soon as the “jackpot prize” was awarded – once out of every million interactions – it was no longer available to be won until “all the tickets” were sold – or every spin needed to award each and every prize was executed. The developer responsible for that type of game changed methods prior to being awarded a license in Great Britain.
AI could handle every element of all of the inner workings of a slot game and even come up with new mechanics that nobody has tried so far. It’s entirely possible that there are labs undertaking the challenge right now. However, the first AI-generated slots were strictly garbage – even though high-quality clones could be created today with the tech.
Garbage in – Garbage Out
In 2019, Ars Technica reported on a duo of developers who created a robot to create mobile slot games in 2013. The experiment created over 1,500 junk games with minor variations in theme and changed perhaps one image to match the theme. The games, available on Google Play, are all dead now and the duo credits their junk with helping “train” Google algorithms to recognize and remove certain types of apps.
According to Ars: “In their GDC presentation, <the devs> mostly left the data up for people to peruse in a laughing manner, but also with a human angle about the costs that might arise if you chase such a robo-generated app dream with profit, not laughs, as the goal.”
The direction today’s developers are taking in an increasingly competitive marketplace seems much better – the images are stunning and on point and the game maths and mechanics are still, presumably, human creations. Where AI-assisted game development goes from here is anybody’s guess.
Source: Here’s How Casinos Are Using AI, Medium, May 31, 2021