Artificial intelligence has made a lot of headlines in recent years. It’s been billed as having the potential to transform the way we do communicate and do business, reduce company overheads, and help create better-looking and better-performing software.
However, it’s also been criticized for its unauthorized use of intellectual property and copyrighted materials, and there are concerns that it will cost jobs within the gaming industry, especially those of creative talents.
AI In Gaming
Despite recent struggles, the gaming industry is one of the largest entertainment industries in the world, with more annual revenue than movies and TV put together. It spans everything from PC to mobile gaming and from AAA premium titles to online gambling, where it is used to create in-game assets and deliver fair gaming at real-money online casinos and sweepstakes-style social casinos alike (source: progressivesweepslots.com).
Common Uses For AI In Gaming
Gaming is at the forefront of technology. Many new developments are driven by consumer demand for improved gaming hardware or software. Artificial Intelligence has already been deployed across the gaming sphere and has been, at least in some small ways, for many years.
More recently, though, as AI has advanced and more companies, educational institutions, and other groups are getting involved in AI development, it is seeing even greater deployment.
However, as well as being praised for its ability to speed up development and improve the player experience, it has come under criticism from some parties.
AI NPC Behavior
We’ve talked about AI in non-player character behavior for centuries. In truth, early examples, such as those experienced in games like Metal Gear Solid, weren’t truly artificial intelligence.
NPCs were programmed to recognize activity, such as specific movements, and then perform a prescribed reaction to that movement. Routines were still scripted, but they used more complex algorithms than they had previously.
This meant that if an NPC in a shooter was below a certain health threshold, they may try to heal depending on the player’s actions. If they were taking continuous fire, they would seek cover. If the player took a certain route or performed a specific action, the NPC would be given another possible action.
Algorithms would determine which specific action to take, and the decision would loop back and start again using the same criteria but different conditions.
Modern NPC AI doesn’t use this looped algorithmic technique. They use behavior trees, rather than loops, where actions can be assigned a specific point value based on a host of criteria, and then the action with the highest points is performed.
Dynamic Difficulties
And AI doesn’t necessarily have to employ the action with the highest points. It can base the decision on player actions and preferences, as well as factors like difficulty settings.
Sony submitted a patent application for dynamic difficulty setting. This enables AI to measure the performance of a player and then scale the difficulty of different gaming aspects accordingly. And that doesn’t just count for overall performance.
If a player is failing most head shots, AI can increase the size of the headshot box, or it could allow upper body shots to reward critical hits. If the player is dispatching regular run-of-the-mill enemies with ease, but struggling with bosses, the AI could change the settings of these specific in-game confrontations.
These dynamic AI settings not only help ensure players enjoy their gaming experience without it becoming too difficult, but they also help ensure immersion.
Keep losing at the same section of a game, and you’re quickly reminded that it is a game. Having to keep changing the difficulty settings also has a similarly abject effect on the gaming experience.
Procedural Generation
Another area where we have seen AI deployment already is in procedural content generation.
No Man’s Sky is a space exploration and survival game. It was panned during its launch, arguably because of AI. The game uses procedural generation to create planets, which means its universe has an estimated 18 quintillion planets.
Random number generators and seed numbers are used, typically combined with algorithms and rules to ensure the planets or other items work. Of course, this scope isn’t necessarily a good thing.
No Man’s Sky was initially slammed by critics and players because planet-hopping became quite boring, and there was very little real discernible difference from one planet to the next.
Modern procedural generation is more powerful. AI modelling can determine what works in world generation, and what doesn’t. It can even track what combinations players have seen and try to ensure a good mix of worlds.
Generative AI
One of the reasons procedurally generated content can become boring is a lack of in-game creative assets. Even with a billion planets in a space exploration game, developers would not be able to create unique features for every planet.
Textures, flora, fauna, NPCs, and other in-game assets would need to be repeated, and this means that while every planet might technically be different, they would start to feel very similar. This is especially true when you add in the algorithms needed to prevent game-breaking world creation.
Generative AI has come under fire. Critics state it lacks the character and flair of content created by humans, while professional creatives point to the fact that it is effectively making their work redundant.
But generative AI can be used to create an infinite number of assets, and these assets can be created on the fly using some predetermined rules. Generative AI content can help increase interest in game worlds.
Personalization
One way AI has been deployed across entertainment sectors is in personalization. You see it when you open Netflix, Amazon Prime, Steam, or even your Xbox, and AI has made the discovery engines that these platforms offer much more effective.
Once upon a time, platforms would use keywords or just genres to determine what content to show. Now, AI can determine whether you like or dislike certain genres, whether you typically enjoy games in first-person or third-person, whether you like short or long games, and much more.
It can use this information, as well as information from other players as well as from game makers, journalists, and reviewers, to recommend games that are best suited to your preferences.
This means you find games you’re more likely to enjoy. It also means that software studios and the platforms are more likely to sell their games because they are being shown to players who have a high chance of liking those titles.
Conclusion
AI is an emerging technology, and there is a race to develop leading AI technologies. These include customer-facing software, generative AI, and a host of others.
In the gaming industry, AI is especially commonly used in procedural content generation, as well as in-game personalization and in the personalization of gaming and other recommendations.
Away from the games themselves, players are likely to come across AI when dealing with gaming company customer service, and it is used extensively in marketing, and the technology is deployed in areas gamers may never see – game testing and quality assurance, for example.
Although there is some pushback against the technology, it is generally considered to be improving the industry as a whole.
