Discover the alarming outcomes of using AI in Games. Get to know the Pros and Cons of AI-generated Content in game development.
In today’s world AI needs no introduction. In the past few months, the application of AI in various fields has boomed, which even includes AI in games, and we have been living with it ever since. It won’t be an inappropriate statement to make that most of us have come to know about the potency of AI because of ChatGPT. And over the period, ChatGPT became a showboat for not just the local audience but big tech firms and many other industries. Along with it, we even witnessed several other AI tools coming into the picture that can do pretty much everything done by a visual artist, a developer, a music composer, a writer, a data analyzer, and whatnot. Well, either it can be treated as a replacement for your desk job or else an extra helping hand to get your job done more efficiently! That’s where the concern begins, is it a double-edged sword of innovation? Or it’s paving the way to balance the technology with the human touch?
These concerns go hand in hand with the innovations and pros of AI being used in various industries, and so does the Gaming industry. AI in games has started to kick off recently, and it seems like regulations are already coming by to safeguard the community and tech altogether. Recently, Valve modified its policies to disallow games from publication that uses AI-generated content like, for example, art, 3d models, music, etc. Though AI tools can help you elevate your production process and make your workflow efficient, it also comes with a risk of getting banned from the platforms like Steam. So, in this blog, we will assess if AI in Games is really revolutionizing the Gaming industry or Robbing creativity.
How does AI in Games work?
There are several aspects to how AI in games works, and it is often misunderstood that it solely refers to AI making a whole game of its own. Well, it isn’t a lie that AI also has the capability to make a whole game of its own upon giving correct prompts, but the topic has a broader sense than the given use case.
AI in games helps primarily to automate several aspects of the development cycle, which includes making of 3d models, 2d arts, generating dialogues, cloning voices, making a quest system, and many more. And for an indie game developer, it is a nightmare to handle all these things by yourself in a certain time frame. Apart from Indie Game Developers, small to medium size studios can even take the help of such tools to uplift their production process remarkably. This is where the role of AI tools and AI-generated content kicks in to ease the game development workflow making it super-efficient. Though it might sound convenient and innovative, it comes with severe drawbacks that developers often ignore and don’t care to check. Below are some of the Pitfalls of using AI in Games and AI-generated Content.
Pitfalls of AI in Games
It might sound a bit odd that where AI is in boom and being used as a savior in the industry to ease out several processes, we are a bit worried about the other side of the coin. But, it is worth assessing the risk of using AI-generated content in games before you delve deeper into it. Here are some of the key drawbacks of using AI in game development.
Lack of Originality and Driving Player’s Emotion
It might be proficient enough to use AI to analyze data and patterns, but it is still limited in various aspects, especially AI-powered by language models. Moreover, it often fails to imagine something novel and later pushes from the available trained large datasets and makes up things. This becomes repetitive as it gives output in the form of derived work from the already available data in its library.
When it comes to grabbing the player’s emotion, it fails miserably in understanding it and results in a bland and uninspiring gameplay experience. However, it’s a matter of fact that the AI can be trained well to handle such instances, but it requires strong technical knowledge to tackle this, which is not everyone’s area of expertise.
Unpredictable Gameplay
Making the entire game using an AI can lead to the unpredictable nature of gameplay, with unstable behavioral patterns and triggering unintended bugs. This can further lead players to become frustrated who are seeking a fair and enjoyable game experience. Additionally, it might also get difficult for the developer to modify certain areas of the code generated by AI as it might break the pattern that has been set up, leading to never-ending errors.
Unintended Glitches
AI-generated content may come up with several unforeseen glitches, which could become technically challenging to identify and fix. Moreover, endless games made with AI can become the utmost difficult to debug, leaving the glitches unnoticed and tarnishing the reputation of the game developers.
Ethical Concerns
AI-generated content may sometimes lead to morally questionable scenarios raising ethical concerns about having an impact on players, especially the younger audience. The AI model can become offensive as they are trained as per the real-world data leading to some racist, sexist, and otherwise offensive inappropriate remarks. Though there have been improvements in the filter and discarding such instances, it Is nevertheless a promising solution as the AI keeps learning every moment.
Alternatively, certain concerns may arise on true creativity using AI and later on comparing with developers who have actually done the project on their own without taking any such external help. These queries are alarming for the developers and designers who have now started to believe that AI has replaced them by staking their job.
AI is not always right!
Yes, you heard that right! We often tend to believe computers and the technology around us blindly, and we rely on them more than human integrity. However, it requires no explanation that we had witnessed tons of stories about technology failure and how entirely depending on these techs had taken away life. That being said, AI in games is no different, and It is a fact that even AI can make mistakes and is not perfect yet. To give a strong base to our statement, here’s a small example of how AI like ChatGPT failed to give the correct result for a simple use case.
We asked ChatGPT to provide us with a piece of code that generates a random integer, but the output should be a multiple of 5, where the random integer should be picked up in a range between 0 and 100. This seems simple even for a beginner coder to pick up easily and write a small code in a few minutes. But, it was surprising for an AI to get this simple code wrong when several developers have already shown it in the past through social handles that ChatGPT can generate a full game for you and do complex tasks and write code for AAA games.
Here's what ChatGPT gave us,
Well, things got more interesting when we asked ChatGPT to provide us with the output of the code. And guess what, it gave us output as per the question, but it’s definitely not the ones from the actual code.
If you assess the code thoroughly, then it’s quite evident that the piece of code written by ChatGPT simply returns an integer between 0 and 100, but it never takes care of the output being multiple of 5. So, what goes wrong here? ChatGPT manipulated the output as per the question and somehow framed the situation as if it were the output of the code it returned. It is indeed more concerning than providing wrong information in the first place. Just think about the authenticity of the results generated by ChatGPT if it can manipulate the result defying the facts!
Moving forward, we let ChatGPT know that the answer it provided was wrong, and only then it apologized and give the right way of solving the problem.
Though it’s a matter of fact that the starting window itself clarifies that it can generate wrong results and is limited in various aspects, our motive is to make people aware of these instances while making them understand the dos and don’ts using these tools in professional fields.
Real-World Scenario on AI-generated Content
There have been many cases coming upfront where AI-generated content has taken a bad shape and later on got banned from various platforms.
Previous year in 2022, a game named “AI Dungeon” was found by Google to contain offensive content and was later removed. Not just in Google Play, but even Apple recently removed a game from its App Store with the name “Replika” because of conducting fake conversations using AI. Such cases raise concerns over the genuine creativity and the skills of the developers as this AI-generated content is nothing but a few prompts, some manipulation, and finally, assembling. This somehow has become a horrifying situation for the small developers who have switched roles from making a game from scratch with their own skills to learning prompts and creating a game out of AI. This not only engulfs the true creativity of developers but also leads them to the wrong path of utilizing AI tools in the right way.
However, there also have been cases where AI has completely replaced the job of a voice artist professionally. Recently, in a dialog with Game Developer, voice artist Victoria Atkin revealed how modders replicated her voice using AI. She claims that the AI stole her voice for the character Evie Frye from Assassin’s Creed and has been used commercially. Along with Victoria, there have been several other cases of voice artists coming together in the past for such problems. But the question here is, whom to blame in real? Is it the AI that’s answerable in this scenario or the developers behind it who has made such a tool in the first place?
In the most recent case, Valve has also started to tighten up its policies in order to safeguard its community from inappropriate content generated by AI in games. It further made clear that Steam is not against banning games using AI or AI as a whole, but it would have a strict eye on the AI-generated content, as it has the highest probability of generating copyright content or any such thing which is not particularly allowed on steam as per their developer submission rule. On a shorthand, it means Valve is not allowing any games with AI-Generated assets to be used, as it poses a threat that the generated content is not under the legal ownership or rights of the user and may fall under copyright which is against the developer submission policies of steam. That being said, Steam doubts the authenticity of the data models used by the developers to train the AI., and that’s where Steam has been working to strengthen the policies to support the use of AI in games and moderate the use case on the other hand.
But the above-mentioned cases don’t mean the ban on using AI in games, and rather there needs to be some sort of regulations and moderation governing the use case and application of AI. However, big tech firms like Ubisoft and Unity have already invested a bulk in AI technology, and now it would be interesting to see how AI gets a permanent seat in the Game development Industry.
The Right way to use AI in Games
Now, after hearing all that’s wrong with using AI, we definitely can’t leave you without letting you know the right way of using AI Tools. AI Tools can be immensely effective when used correctly, and nothing should dissuade you from using AI Tools to improve your workflow. And that’s exactly what AI is supposed to do, improve your workflow, not replace it.
People misjudge AI to be a tool that is supposed to replace us and take over all our jobs, and due to this presumption, users often take AI-Generated content as the final output of their requirement, and this couldn't be further from the truth.
AI-Generated content’s primary intended use is to take the tedious tasks out of our workflow and generate quick concepts for all kinds of creative projects. Think of AI as a personalized search tool that gives you accurate references to base your designs on. Imagine searching the web for the perfect references for your project; while you could put in the exact prompts you need for your references, and viola, you can get a nearly unlimited supply of designs, all catered towards your specific goal.
After this, your job is to take the output and finesse that as per the project requirement, which can solely be done by someone who has an eye for that kind of work. Only an avid programmer can find the flaws in an AI-Generated algorithm and optimize it, while a layman would be happy with the result and implement it, only to find out that it doesn't provide the full solution to the problem. Similarly, with art, AI can generate thousands of images but using them directly solves no purpose as none of them are exactly what you need to integrate into your project. The colors might be off, the pose might not be right, and it’s your duty as an artist to take the art asset to the next step.
So, though using AI might not always be helpful in a direct way, you can still effectively use it as an extra helping hand to your work. And it is no doubt that using these AI tools in the right way will boost your productivity!
The employment of AI in games possessing a significant risk that far outweighs its perceived benefit still remains a question to be tackled until a few more years. Though lack of originality, authenticity, potential bugs, and unpredictable nature all contributes to the argument against the use of AI in games, we must not forget that we are in the phase of experimentation, and the technology is not fully matured yet. On the other hand, game developers must understand the importance of prioritizing human integrity and creative expression over relying completely on such new-age tools. That being said, this was our take on what we feel about AI in Games and AI-generated content used in game development. We would love to have your thought on this and bring on new perception to the story.
With Love From VOiD1 Gaming
