Friday, March 8, 2024

GENERATIVE AI COULD HELP VIDEO GAME DEVELOPERS COMPLETE MONTHS OF WORK IN MINUTES. CAN IT SAVE YOUR BUSINESS?

The video game industry needs a hero, and artificial intelligence might just save the day.

While 2023 was widely-regarded as one of the best years ever for gaming, with a record-breaking number of critically acclaimed releases, there were warning signs that the good times wouldn't last. Venture capital firms invested just $4.1 billion into gaming companies in 2023, a far cry from the $14.6 billion invested in 2022 and the $16.2 billion invested in 2021, according to a recent report from the research firm Pitchbook. Now dealing with reduced funding, the $184.3 billion video game industry--like many businesses across sectors--is hoping AI can help it bridge the gap between its ambition and its resources. 

The reduced VC activity "likely represents a more realistic level of investment than the peak years of 2020 to 2022, which attracted non-endemic and 'tourist' investors during Web3 and metaverse hype-cycles," according to the Pitchbook report. Large gaming companies used the influx of funds from 2020 to 2022 to scale up their staff in order to take on hugely ambitious games with budgets topping $300 million. In this suddenly more difficult funding environment, those same companies have been undergoing massive layoffs and are scaling their ambitions back down. Just in the past few weeks, Sony announced it would lay off 900 PlayStation employees, Microsoft said it would lay off 1900 Xbox employees, and Electronic Arts said it would lay off around 670 employees, or 5 percent of its workforce. 

"The party is over," according to Ilya Eremeev, co-founder of gaming-focused VC firm The Games Fund. "Now it's time for the hangover." Eremeev says that these inflated budgets can "make your margins so thin that even games with great reviews and sales can still be unprofitable, because the development cost is so high." 

So, how are these developers supposed to execute on their still-ambitious visions with diminished resources? The answer, at least according to a few major players in the gaming space, may very well be generative AI. In November, Japanese gaming titan and Final Fantasy creator Square Enix announced an investment in Atlas AI, a startup providing game developers with genAI-powered tools for the creation of 3D assets and models, essentially the "stuff" that makes up a digital world, like buildings, landscapes, and props. Atlas scans reference images like concept art and mood boards to understand the aesthetic a designer is attempting to achieve, and then renders a "catalog" of 3D assets that can be further customized by artists. 

Atlas' founder, Ben James, a former architect, started the company with an ambition to create a tool that could scan architectural plans and blueprints, and instantly render a high-quality 3D version of the project. But in speaking to representatives from various industries that were interested in the tool, it became clear to James that "the scale and speed at which game developers build necessitated this type of technology." 

The amount of computation needed to create huge-scale game worlds with hyper-realistic graphics has increased exponentially over the past decade, according to James. In addition to being expensive and requiring huge amounts of manpower, designing these worlds is a lengthy process that often leaves little time for revisions. "If you've spent six months developing a particular style for your game" says James, "you're going to be pretty hesitant to leave that style. With AI, you can change those style references and generate a whole new world in minutes. It's going to allow developers to be way more flexible in the early stages of the design process." 

"In recent years, the rising cost of game development has been challenging for the entire game industry, and gen AI is expected to help streamline the process in new and exciting ways," Square Enix general manager Hidekai Uehara said in a statement. "We are excited about Atlas' unique technology and look forward to seeing how it might unlock efficiencies in our business."

Kylan Gibbs, CEO and co-founder of Inworld, a startup developing generative AI-powered tools for the creation of non-playable characters and story elements in games, shares Uehara's belief that AI can be used to streamline existing development flows. In November, Inworld announced a sweeping partnership with Xbox to develop a generative AI-powered tool that would speed up the process of narrative development in Xbox games. According to Gibbs, most game writers at major studios have to go through dozens of steps before they get a sense of if an idea they had months ago is actually any fun to play. With the generative AI tools being built by Inworld, Gibbs says writers could create a playable demo for internal use "in minutes instead of months." 

It's not just big gaming companies that are getting in on the AI action, as smaller studios are also making clever use of more consumer-friendly generative AI to drive efficiency. Rob Lester, chief creative officer at Michigan-based independent game studio Zollpa, has used ChatGPT to speed up his team's quality assurance and playtesting process, essentially the process of poking and prodding the game to find bugs and address gameplay issues. 

Unlike large studios, which often employ hundreds of contractors to perform QA services for their games, Zollpa's team of eight conducts QA and playtesting entirely in-house. For Zollpa's latest game, an online arena shooter called RoboSquad Revolution, Lester had the idea to record the team's QA sessions, upload a transcription of the audio to ChatGPT, and ask the chatbot to analyze the document, noting any instances in which bugs or issues were mentioned. Those notes were converted into action items, which were automatically assigned to team members based on their expertise. "It streamlines the process of gathering the information we need to act on tremendously," says Lester. "Plus, it's pretty hard to play a video game and take notes at the same time, basically impossible."

Gibbs says we're still a few years away from the release of any major games built to take full advantage of genAI, but outside of gaming, many businesses are much further along when it comes to harnessing AI in meaningful ways. Last week, Klarna announced that its new OpenAI-powered customer service chatbot was doing the equivalent work of 700 employees. In February, Adobe released an AI assistant that can read long documents and answer questions about their content. And a recent survey of companies that rely heavily on AI found that some have seen so many productivity gains that they're considering instituting a four-day work week.

No comments: