Heba Al-Sayed wrote Monday, April 1, 2024 07:00 AM
It is not surprising that artificial intelligence It’s the biggest topic of discussion at this year’s Game Developers Conference, where emerging technology was spread throughout the conference’s exhibit hall, with huge booths showcasing demos from companies like Inworld.
During the event Nvidia showed off some of the most impressive tools, an experience created by Ubisoft using Nvidia’s Ace microservice that caught the eye, as it allowed players to speak to NPCs through the microphone and quickly get convincing AI responses.
During the event’s annual Game Developers’ Choice Awards, presenters railed against the fast-approaching wave of generative AI that poses an existential threat to human creators.
However, not every use of technology is aimed at taking humans out of the creative process, some, like the AI tools that power Candy Crush Saga, make a stronger case for AI use cases that are more beneficial to creators.
Developers at King, which was recently acquired by Xbox, are also revealing the technology that powers its games, with one panel in particular focusing on how the studio uses artificial intelligence to create Candy Crush Saga levels.
“It painted a more human perspective on how the company uses the tool, where we apply AI as a utility to provide solutions for playtesting,” said Sahar Assadi, Director of AI Labs at King.
As for the level a designer creates, Assadi explained, “We get data about what the gameplay looks like and a lot of metrics about how challenging it is, which helps the designers shorten the mission cycle.”
According to Assadi’s explanation, Candy Crush’s AI tools are all about saving time, which is essential for a game like this, where the mobile megahit has consistently kept players engaged over the course of 12 years due to its astronomical number of levels.
At the moment, it has over 16,000 levels, and this creates a huge amount of work, as each level needs to be properly balanced and play-tested to make sure it is fun and not too difficult.
For King, this was a time-consuming process that left a lot of room for error, which AI would eliminate.
“Historically, it’s been common for developers to do one or two rounds of opening a level, playing it, getting to know it, and then, whether they’re happy or not, several iterations happen,” Al-Assadi said. “That used to be the most time-consuming part but now they click and run the test run over 1,000 times to get more accurate insights, but also to get additional metrics.
The criteria relate to the questions of whether it is playable? How challenging is the player? Then the designer gets all of this vision and decides which direction to improve it, and that comes in a matter of minutes.
This massive amount of digital playtesting is crucial for an engagement game like Candy Crush Saga, and free-to-play games like this can easily attract a large number of players, but they can just as easily lose them.
