ART-ificial Intelligence in Gaming Part 3: Implications and Conclusion
GAN = creative AI = ?
This is part three of a miniseries, so please read parts one and two before continuing:
In the other parts of this miniseries, we talked about how GANs are used to generate new video game levels that adhere to all the rules and game physics. There are endless possibilities of how it can revolutionize modern gaming, including having more realistic graphics which can truly enhance the experience of virtual reality.
If you think about it, human intelligence is creativity. The “smartest” people you know are the ones who think outside the box; they think creatively and can make connections. Therefore, true artificial intelligence must have the capability of thinking independently and creatively.
GAN = creative?
Since GANs are used to generate completely new data that could plausibly be from the original dataset, it’s essentially “creative” AI. Let’s say we feed the model a diverse dataset of illustrations, from Picasso paintings to the digital art on TikTok and everything in between. Results would most likely be unique mixtures of the art styles — and isn’t that what we call “creative” or “original”?
Yep! Artists create their unique art styles by mimicking aspects of certain things they see. In the video game examples, the AI would mimic parts of existing levels. What other implications does creative AI have?
Benefits of Creative AI
Beyond helping game developers create more compelling video games, GANs could learn the rules of the real world much like how they learned the rules of Pac-Man. They could be used to program robots that would easily integrate into society, recognizing things such as social norms and laws through observation.
In general, having more creative — or more intelligent — AI benefits pretty much any field of interest you can think of. Some that I came up with:
- more unique video game bots and terrains
- more conversational or helpful chat-bots
- cool, modern art and music made by AI
- environment-adaptable robots
Of course, this comes with some major drawbacks :(
Drawbacks of Creative AI
The more we invest in researching how to create better AI algorithms, the more dangerous it becomes. Who knows what will happen when robots scheme to take over the world? Just kidding… Unless…?
Jokes aside, some people may be worried about creative robots stealing our jobs. But really, data privacy and security are the most pressing issues in developing AI because to have more effective algorithms, more data needs to be collected and fed into the system. This creates all sorts of problems when people don’t want their every move to be tracked and potentially be used against them.
Deepfakes are fake videos or images of people or events that seem real, created using GANs. Politicians are often victims of deepfakes because they are manipulated to appear to say something that they did not. If you’d like to learn more about deepfakes, I recommend checking out George Dvorsky’s article.
CONCLUSION
This is the end of my first miniseries and the last post for 2020! Throughout the miniseries, I talked about how AI is used in video game graphics, namely the generation of new stages or terrains. Then I went into how the AI is built (GANs), and finally concluded with implications beyond the virtual world. It’s up to you to decide if the benefits of advancing AI outweigh the drawbacks. I will be writing more about AI next year, so follow me on Medium (or don’t).
See you soon :)
Danielle Trinh is a Student Ambassador in the Inspirit AI Student Ambassadors Program. Inspirit AI is a pre-collegiate enrichment program that exposes curious high school students globally to AI through live online classes. Learn more at https://www.inspiritai.com/.
