The presence of artificial intelligence (AI) is starting to appear at my workplace.
This has brought two sides into focus. The first is a question that those involved in the academic part of the business are asking themselves right now: how do we ensure out students are using AI appropriately to aid their studies, without becoming reliant on it to complete their work?
The second is a subject closer to my role within the IT department, where we want to find ways to empower our users with new technology and automate repetitive processes. For instance, AI has recently been integrated into our main IT service management (ITSM) system to help analysts send more customer-friendly updates related to support requests.
But what’s the link to gaming? This isn’t the best introduction I’ve ever written for a post but I promise there’s a point. Over the past several years, I’ve supported a number of Kickstarter campaigns which have primarily been for various types of games. I’ve noticed a gradual rise in detective-themed projects during this time – take the recent Scene Investigators by EQ Studios as an example.
It was only a matter of time before campaigns that merge the detective theme with AI technology started making appearance on the crowdfunding platform. Although I’ve seen a few of these now, each repeatedly claims to be ‘groundbreaking’ or ‘the first of its kind’. Such projects often feature partner or suspect characters powered by ‘cutting-edge technology’, with developers aiming to create immersive experiences filled with ‘convincing dialogues’.
But as appealing as these pitches might sound, it doesn’t guarantee success with backers. For instance, AI Suspects: Unmask AI Crime by RoleCraft AI has only managed to collect £317 towards its £8,226 goal in its first week (at the time of writing). CaseDog Interactive fell short last month, securing less than 12% of its £6,000 target. And then there was Vaudeville by Bumblebee Studios, which raised only £510 from 23 backers when it was launched on Kickstarter in March.
Despite the failure of that last campaign, the game made it into early access on Steam in June. It promotes itself as an ‘experimental whodunnit game based on the latest AI technologies to generate dialogues in real time’. Playing as the famous private investigator Detective Martini, the objective is to use your wit and intuition to uncover intricate secrets lurking within the city’s vibrant streets and shadowy corners. Recent reviews are mixed, with only 60% currently being positive.
Players have reported various issues, such as characters ceasing to respond after being asked an initial question. The AI’s inability to recall information from previous conversations throughout the gameplay results in frequent random topic changes during dialogue and this completely breaks the immersion. Response times are also problematic; it obviously isn’t desirable to have an investigation mechanic which requires you to wait for ages before a suspect replies.
Inworld Origins by Inworld AI has fared somewhat better since its release on Steam, perhaps due to it being a free title. I vaguely remember seeing this game several months ago and skipping over it because of poor reviews. However, it now has a recent ‘very positive’ score – though most of the comments seem to have been left for comedic effect. ‘Convinced everyone I was Barack Obama’ and ‘called me Space Cowboy and proceeded to make yeehaw references’ are the kind of reviews you can expect.
Players have reported issues similar to those mentioned for Vaudeville. The AI struggles to understand accents which aren’t generic, leading to frequent misunderstandings by the characters when the voice function is used. There are also problems with suspects abruptly falling silent after a few minutes, along with numerous crashes. Given that this was meant to be a title to serve as a case study to show how AI non-player characters (NPCs) can transform game narratives and experiences, it’s safe to say it isn’t a shining example.
So, why are video games that incorporate AI in this manner not being well received? The most obvious answer based on the Steam releases highlighted above is that the technology isn’t yet capable of providing an immersive experience. I’m old enough to remember when virtual reality (VR) first emerged in gaming in the early 1990s. Just like then, AI isn’t yet advanced enough and needs a few more years to develop. Perhaps we might start seeing NPCs that feel like real people at some point in the future.
But there’s a deeper layer to this situation. It’s not just published titles which aren’t doing very well – even future games on Kickstarter, still in the development phase, are failing to generate enough interest to meet their funding targets. There’s a possibility that the entire release chain is being adversely affected by our general aversion to the use of AI in video games. Last week, I reached out to some friends to talk about this and hear their thoughts on the matter.
While they all held different views, the common thread among them was that people generally feel suspicion toward the technology. They see it as a tool exploited by unscrupulous corporations who want to maximise profit by replacing human talent with robots. As one friend pointed out: “This is already happening in a lot of creative industries and it absolutely sucks. This also means the general sentiment towards AI is that it’s a horrible job-stealing monster, so any game that has AI as a selling point is going to get avoided or review bombed.”
However, this doesn’t mean they’ve completely ruled out backing or playing video games which feature AI. While one friend hesitates right now because ‘the tools available are still primitive’, the others expressed openness to the idea as long as the game aligned with their interests. As someone eloquently put it: “Technically, games have had AI for a long time, plus I think AI-generated games can be an interesting space, so long as it’s not the entire game that’s AI-generated.”
The key takeaway was that AI will likely remain a niche element for some time. While we may come across a few quick money-grabs on Steam, the technology is destined to play a more substantial role in all industries in coming years. We might even stop referring to AI altogether as it gradually becomes an integral part of everything we consume. There was one particular comment which stood out: “AI and game development technically go hand in hand, but it’ll take the curation of a creative human to leverage it and not over-use it.”
As the technology evolves and developers find the right balance between inventiveness and AI assistance, there’s a chance we’ll move beyond our initial distrust. Perhaps we’ll become more open to the idea of AI-generated content as long as it enhances, rather than replaces, human creativity. Although we’re currently cautious and sceptical, the future holds great potential for AI to transform the way we play and experience video games.