Artificial intelligence fascinates and worries at the same time, upsetting our perception of technology and its capabilities. John McCarthy, one of the pioneers in the field of AI, described it in the last century as “the science and engineering of making intelligent machines”.Used wrongly and especially indiscriminately, the concept brings together above all a set of algorithms allowing to imitate human thought, capable of creating, correcting itself, learning and anticipating a situation. While the use of AI in predictive and generative algorithms is slowly starting to make its way into our daily lives, video games are not immune to the wave.
What is AI?
Today, AI is used as a catch-all term that encompasses both real intelligent proposals and simple automation mechanisms. However, artificial intelligence goes far beyond the action-reaction system. The field can be divided into several fields of action, with generative AI on one side, cognitive AI on the other, and other categories. All of these forms are currently in vogue for everything they bring to the entertainment sector. ChatGPT, Midjourney and Dall-E are just a few examples of the most widely used and, above all, most visible software.
But AI is also a huge submerged part, which insinuates itself into programs in the form of digital tools that adapt to human activities... as another human would. These are personalized recommendations, voice assistants, or even facial recognition software. We can no longer list all the uses of artificial intelligence in everyday life, via tools that are as invisible as they are useful.
The situation gets complicated when we talk about art, the only aspect of creative genius that humans do not concede to their artificial counterparts. Video games, as a form of digital art, are not immune to current issues concerning ethics, but also the relevance of using AI as a creative tool.
For the worst…
During the Game Developers Conference last March, Ubisoft unveiled its Neo NPC (understand the non-player characters of a new kind). This technology intended to revolutionize the way players interact with their in-game entourage by including intelligent NPCs, capable of adapting to the chapter of the story in which they find themselves (in the case of open worlds for example) or of adapting their speech according to the player's response (in the case of multiple-choice dialogues).
To do this, the studio formalized a partnership with two big names in the industry, namely Nvidia and InworldAI. But it didn't wait for this event to bet on artificial intelligence. Already in March 2023, Ubisoft announced the creation of a tool called Ghostwriter that made it possible to generate NPC dialogues in the background. We're not necessarily talking about those that are important for quests and story, but especially those that populate the open and vast worlds, that make them more human and realistic.
These two generative AI tools are not the only proposals to have emerged in the video game sector. One game in particular stood out from the crowd, being almost entirely designed using AI. Developed by the Finnish Jussi-Petteri Kemppainen, this futuristic point and click with the appearance of a prototype would have been born from the synergy of two generative image AI tools: MidJourney and Stable Diffusion, capable of creating visuals from scratch from text prompts.
The result is stunning. Not only are the images beautiful, but the game's visual identity is clearly identifiable, even though the universe created only exists from already existing models. Although he put everything together, the developer did not call on any artist to obtain an effective graphic direction. For fear of a major digital replacement or an uncontrollable acceleration of artificial intelligence, criticism is multiplying, and players are only slightly receptive to the approach.
Ubisoft is the perfect example. With each new release of a “revolutionary” tool (in the wind of what is being done at the moment), negative comments come in waves, between pure and hard hatred, fear and disgust. It is clear that the ordinary gamer adores generative AI for personal and creative use, but abhors it when it comes to professional use.
… and the best part?
In reality, AI is already everywhere in video games. In the form of an assistance tool rather than an idea generator, artificial intelligence goes unnoticed and is even acclaimed for its relevance and efficiency. It interferes in the moderation of communication sites, in chatbots or even in the user interfaces of consoles and other gaming-related applications.
Twitch, a live streaming platform owned by Amazon, is one of the most present players in the representation of video game content on the Internet. The service is mainly used to share everyone's passions with the general public, often with respect and kindness. But Twitch is no exception to hateful and offensive behavior, so moderation is extremely important to preserve the integrity of each user.
In 2021, the platform unveiled a brand new tool called "Suspicious User Detection" based on artificial intelligence. The tool helps fight cyberbullying by identifying accounts that try to circumvent banning rules, for example by creating multiple accounts. The news did not cause as much of a stir as any ChatGPT update, and yet the tool is still used on a daily basis.
Recently, it was the future PS5 Pro that made headlines and its relationship with artificial intelligence. According to rumors, Sony's next console, or the one after that, should integrate a dedicated processor that would boost the console's performance. The AI upscaling technology, described as an equivalent to Nvidia's DLSS, would allow games to be displayed in 4K and 120 fps. Similarly, the machine would soon be able to play you “in your place” by using techniques for analyzing and reproducing your in-game behavior. However, all of this remains optional, and will be integrated into the already very complete PlayStation ecosystem without intelligence becoming the console’s main selling point.
Criticized from then on that it attacks artistic fields, AI manages to be trivialized, even praised in cases where it tends to be invisible. That artificial intelligence fascinates is rather logical, but the fact is that it also worries. Generally speaking, we see that all innovations worry before being accepted.
Ready-to-use artificial intelligence
When it comes to video games, what worries consumers the most remains the status of developers within the industry. The sector is going through troubled times. Since 2022, studios, publishers and manufacturers have been laying off their employees left, right and centre, taking the hit from a prolific pandemic and a brutal return to normal. Last year, 10,500 jobs were eliminated, while 2024 already records 10,000 layoffs in less than 6 months. The trend is accelerating and this does not bode well for the future of professional video game activity.
The shadowy threat of terrible artificial intelligence is an additional sword of Damocles hanging over the heads of developers. Faster, less expensive and not subject to social rights, technology only counts on the good faith and goodwill of companies not to replace humans in the most difficult and/or creative tasks, or those that require the most financial investment.
SAG-AFTRA (Screen Actors Guild-American Federation of Television and Radio Artists) also took up the subject just a few months ago. The association that historically defends staff from the audiovisual world took advantage of the demands in Hollywood to address the use of AI in video games. Its goal was never to ban artificial intelligence, but rather to regulate its use so that motion capture or dubbing actors for video games work in a healthy environment, without fear of having their image/voice/career stolen.
Legislating rather than repressing is the option that SAG-AFTRA has chosen and it is the one that led the association to pass an unprecedented step with Replica. The company is authorized to create voice doubles of unionized actors and provide them to video game studios, in a “fair and equitable” environment for dubbing actors, who will finally be able to reap the financial benefits. In this context, companies using talent through Replica, and therefore going through the protection of SAG-AFTRA, know that digital and vocal clones will be based on real people, having given their consent for the use of their image, and that each use will be controlled.
The agreement was not well received by all members of the movement, but it at least had the merit of laying the foundations for regulation. SAG-AFTRA preferred to join the transition movement rather than swim against the current of a technological phenomenon that is already taking hold everywhere without us seeing it.
Can legislation ensure the future of AI?
Organizations and companies are not waiting for the general public to decide on the fate of artificial intelligence to seize it in all possible ways – more or less ethical. It is therefore more than necessary to have legislation that answers all the questions that we legitimately ask ourselves about how AI works and especially what it feeds on.
Today, even the most advanced generative AIs go through extremely large databases to assemble images, colors, shapes or ideas, and thus rebuild something new from many existing things. Who should we credit then when Midjourney produces an image from a prompt, which will go and draw from databases where there are thousands of works by very real artists? If this question is still thorny, that of the vocal clones used by Replica will at least have the advantage of having a clear answer, thanks to the agreement reached by SAG-AFTRA.
This is a crucial step towards a more ethical adoption of this technology in industries as large as video games. However, the question of the impact of AI on employment and human creativity remains. It is essential to continue legislating and debating these issues to find a balance. The future of AI will depend on our ability to integrate it responsibly and fairly.
0 Comments