For most people, the gaming industry seems pretty secure against the rising threat of automation. It’s a creative industry that requires developers to design entire worlds, characters, and storylines.
Yet, this isn’t actually the case. Automation is already changing the gaming industry, and will continue to affect the jobs market for some time.
Automation and Jobs
This is not a phenomenon limited to gaming. Automation is changing, or threatening to change, practically every single industry in the world. Elon Musk’s Tesla has been developing automated trucks that could be used to transport cargo across countries without the need for a driver. This could see up to 3.5 million truckers lose their jobs in the United States alone, with many more at risk in other countries.
Not all jobs will be affected equally by the increasing levels of automation. For example, experts say the likelihood of driving and post office roles being replaced by technology is almost a 100% certainty. However, jobs that require personal interaction, like airline cabin crew and nurses, are at much less risk.
As a general rule, jobs that require repeated manual work are likely to succumb to automation much quicker than those that require creativity and building relationships.
Automation in Gaming
Given that creative jobs should be at less risk, it is perhaps surprising that video games are already, at least in part, designed by artificial intelligence. Large open worlds and maps can already be partially created by using AI. This means that even small development companies can produce large-scale games.
For example, the British company Hello Games had a team of just 20 game developers in 2018, and only four in 2013. Yet, it has been able to develop the massive exploration-survival game, No Man’s Sky. To create such a large open-world game, the company needed artificial intelligence to create a “procedurally generated universe”. No Man’s Sky uses algorithms to generate environments that mimic patterns found in nature and manipulate graphical elements that have been created by human designers.
The game Diablo also uses these artificial intelligence techniques to create maps that vary each time you play them. Ever since the first version of the game was released in 1997, it has used randomization to change the layout of the map, the items found in it, and the quantity and placement of monsters. This defining feature made it a fan favorite and helped Diablo to develop a cult following.
For game design though, automation, at least in its present form, is a tool to be used by developers, rather than something that will replace them. With video games becoming larger and more detailed, automation is the only way that games can be developed quickly enough. This could mean fewer people are needed, but it won’t mean automation will completely replace video game designers.
Quality Assurance and Testing
The most time consuming and repetitive element in the production of video games is quality assurance and testing. Publishers employ hundreds of people to test every single element of a video game to ensure it behaves as intended and to find bugs.
Without stringent testing, games can crash on players, creating frustration and even resentment that can negatively affect the commercial success of the game. Driv3r for PlayStation 2 was widely criticized for this, and at the time downloadable patches were not available, making the game unplayable.
Automation in software testing has been common for some years, helping to speed up the production of apps and computer software. For video games, though, the testing process is a lot more complicated, so automated testing wasn’t possible.
Artificial intelligence is changing this though, as it can recognize objects in a similar way to humans, whereas procedural automation tools can’t do this. AI testing doesn’t need to have hard coded steps, so it can dynamically adapt to changes in the game.
While AI testing still has a way to go, it’s getting closer to rivaling human testing. This could be where the biggest job loss to automation occurs. It’s likely jobs will still be needed to code, monitor and manage the AI that does the testing, but some roles will still likely be lost.
