The Japan Times - AI revolution in video games has industry players treading warily

But even the ultra-connected industry is eyeing the innovation warily, with fears growing that jobs could be made redundant and artistic creations usurped.

"AI is really a turning point," according to Julien Millet, an AI engineer and founder of United Bits Game studio, who attended the industry fair this week.

Responsive non-playable characters or the automatic generation of images, code and game scenarios are among the possible uses for developers using AI.

AI is also capable of instantly producing illustrations from text, allowing producers to better "transmit their vision", according to Millet.

But the images dreamed up by AI could threaten the work of concept artists, who visualise the video game world before it is created digitally.

"I am worried for those jobs," Millet said.

- 'Dream world' -

Attracting tens of thousands of video game lovers every year, Gamescom is an opportunity for studios to showcase their latest creations.

Many gamers turn up in cosplay costumes, as they throng the stands to try out the potential new hits -- this year including some that prominently feature AI.

Club Koala from the Singaporean studio Play for Fun offers players the chance to "create their own dream world, a personalised paradise island with unique characters" generated using AI.

"AI has become an integral part of everyday life" and has a "huge potential to take the gaming industry to the next level", Play for Fun CEO Fang Han said in a statement.

Berlin-based Ivy Juice Games also said it now uses AI through its game creation process.

"We use it to generate lines of text... to get some more storytelling into the game," Linus Gaertig from Ivy Juice Games told AFP at Gamescom.

It is also using AI "to generate code", Gaertig said, offering a new way for developers to build the games themselves.

"(AI) makes the game more unpredictable and so the game feels more real," said Sarah Brin of Kythera AI, which uses the technology to generate character movements.

A case in point was demonstrated by US chipmaker Nvidia when it introduced the world to ACE, a software aimed at developers to create "intelligent in-game characters" using AI.

In its promotional video for ACE, a player speaking through a microphone is depicted having a conversation with a virtual ramen noodle chef in a sci-fi bar.

How is the chef? "Not so good," comes the answer -- crime is on the rise locally and the chef is worried.

But using AI to create sprawling virtual worlds could clash with claims to the intellectual property rights on the original images used to produce them.

"If you are a major publisher and then you use generative AI, turns out what you've used infringed on some copyright, then you're open to some vulnerability there," said Brin from Kythera AI.

Unlike many of its competitors, Brin's company has decided not to train its AI on open databases.

After all, in the US, artists have already jointly launched a suit against Midjourney, Stable diffusion and DreamUp, three AI models created using images harvested from the internet.

S.Yamada--JT