Will AI Replace Games Developers? New AI Replicates the Mechanics of a Video Game

By  | 


Isaac Asimov’s ominous observations back in 1942 no longer look as farfetched and fictional as they once did due to the increasing power of AI. Robotics is growing at a steady rate – expected to have a market value of $226.2billion by 2021 – and the way humans are beginning to discover how AI can be used to our advantage is leading the way in determining the technology. Still, the use of AI throws some interesting philosophical and practical issues in the air: will AI replace the human workforce? Will creative jobs and professions that AI can’t replace become even more saturated? It was recently reported that AI technology is being used to develop games – which was once considered to be out of their remit of capabilities due to the creative elements involved. Will the future of gaming be in the hands of robots?

Robots Working with Robots

An AI developed at the Georgia Institute of Technology that was able to replicate the mechanics of a game just by watching it being played. The technology is aimed at increasing the rate at which games can be developed – by giving AI units the legwork to do while developers work on the rest of the game. Essentially, an AI system will be working on developing the workings of another AI system. Supercomputers have indeed shown us that they can work through a vast amount of data, such as the Facebook algorithms that can store information about us and use it to inform the internet what to show us.

Deep Blue vs Kasparov

Measuring the smarts of a robot is a complex procedure – akin to discovering how intelligent animals are. One way scientists can do this is to pit the robots against games that require mental dexterity, forward planning and strategic thinking, and reactiveness to an opponent’s move. The famous supercomputer Deep Blue beat Kasparov at chess in 1997, giving us the indication that one day AI may become as intelligent as we are. Indeed, it is argued that big data was spawned from the IBM technology that helped Deep Blue and that by working out every single iteration of chess matches, a supercomputer would always have the upper hand.

Left Brain vs Right Brain?

Chess, however, is a game of complete information, where anyone playing, AI or human, is called to calculate as many steps ahead as possible. What happens when the element of chance comes into the mix? Blackjack requires a fair amount of problem-solving, especially when playing with others – and AI was proven to be able to logically figure out how to win the game. The supercomputer technology allows every possible combination of blackjack card draws to be analyzed much quicker than a human physically could. This shows us that AI is beginning to think like us – just as we programmed them to. Yet, these examples merely show successes in areas that require logic, chance, and strategy which are left-brain analytical themes. AI has yet to showcase talents in right-brain creativity or empathy pursuits, which may require games to have a mixed human/AI team of creators.

Harry Potter and the Artificial Intelligence

A lot of games focus on the narrative that binds the gameplay together, something that is regarded as important by gamers. So for an AI to truly create a playable game, as well as mastering the physics of the gameplay, they also need to have a grasp of how to frame the gameplay. One researcher conducted an experiment using Harry Potter novels. In 2016, the deep learning algorithm was given the first four books of the $25 billion franchise and was asked to produce a similar chapter. The resulting chapter, which Max Deutsch published on Medium.com showed that while AI can grasp some aspects of the fiction, they miss the mark massively – at least currently. The story uses words Rowling may have used, and sentence structures found in the books but made literally no sense to read. So, while AI may be able to develop some aspects of gameplay, game developers who can think of the overall narrative, mechanics, and plot need not worry about AI taking their jobs.

So, What Does the Future Hold?

Games development in 2017 is a complex procedure. While there is a great deal of inner game mechanics that AI is beginning to assist with – much like a factory conveyor belt system working on batch producing parts that were once made by hand – it is likely a long way off before AI can harvest autonomous thought to create a narrative for the game. It’s likely that given the changeable nature of the future of games – such as the development of AR and VR, AI maybe even further off. Artificial intelligence could probably create the kind of simplistic game that works in an arcade, but when it comes to the rich gameplay of today’s gaming giants, a human touch is definitely needed in order to give the fans what they expect of gameplay today. The AI just helps speed up the processes that they can do.


You must be logged in to post a comment Login