What AiMation Studio’s AI reality show “Non Player Combat” says about race

One AI-centric studio has released a reality TV series on YouTube, where its characters must battle to the death.

Imagine a reality TV show where six extremely attractive people battle to the death to secure 500,000 coins.

Think of it like the Hunger Games, without the backdrop of a dystopian, socio-economic crisis, but with the same premise: one person will emerge victorious, clutching a grand prize.

However, instead of real people, it features AI-generated characters navigating their way through an artificial, dystopian island.

In Non Player Combat (NPC), perhaps a play on the term “non-playable characters,” which is also a form of AI, six “contestants” are flown to various habitats.

One player is transported to the Arctic landscape, another to the jungle, another to a dystopian wasteland, and another to the desert.

Each character must try to survive in their environment, which is plagued by dangerous animals, a lack of resources such as clean drinking water and food, while also avoiding being killed by their opponents.

The six contestants, Travis, Ed, Madison, Micah, Eliza, and Kai, all have different backstories that justify their reasoning for entering the competition.

Eliza is an ex-inmate who was locked up for eight years after “hurting someone close to (her),” while the other AIs are survivalists and ex-army veterans.

While AiMation Studios, the creators behind NPC, did a good job at exploring diversity amongst their characters, I did notice while watching that stereotypes and biases are perpetuated throughout the storytelling.

For example, Kai Chen appears to be of Asian descent, while remaining relatively ethnically ambiguous.

This is evidenced by one shot where Chen is standing in the middle of what looks to be a subway station with other Asian AIs hurrying around him.

But what I found troubling was the stereotypes perpetuated by AiMation Studios.

Instead of showing where Chen grew up, his family, or his education, what the company decided to show was the only Asian character practicing martial arts.

“My family pushed me hard growing up, and I honestly think that made me resilient in a way that I don’t think most people are,” Chen’s character recounts.

This is another example of stereotyping or potential bias, as it perpetuates the idea that Asian families are super strict and tough on their kids.

This was the main bias I saw while watching the first episode of NPC. However, the show does set up events to come with the mauling of Travis by a polar bear in the Arctic, and when Ed gets bitten by a snake in the desert.

While Travis gets off relatively lightly, as the polar bear, for some reason, decides to spare the AI from being eaten alive, something that certainly wouldn’t happen in real life, Ed doesn’t seem to have the same luxury.

As the only black character in NPC, I found it uncomfortable to watch Ed vomit blood and enter a state of paralysis after being bitten by a snake.

In the real world, if you’re bitten by a venomous snake, your chances of survival vary depending on the type of snake, whether you’re in a developed country, and whether you have access to medical care.

However, Ed is in the middle of the desert with no resources and no medical care, which makes his chance of survival relatively bleak, considering the AI is throwing up blood.

This made me think of the common trope in thriller and particularly horror movies in the 60s, 70s, and 80s, that black characters were often killed off first.

Yet, I’m reminded while watching this 38-minute episode that I should probably suspend my disbelief and just take NPC at face value.

After all, it is just AI slop.

AI is biased. Period

AI models mirror human biases. That’s kind of a fact we all must reconcile with.

Large language models like ChatGPT, Gemini, and all other LLMs are trained on information generated by humans. We, as humans, are naturally biased creatures.

Therefore, AI is trained on biased information, which is then regurgitated and perpetuated.

What’s unfortunate is that people may perceive the words of ChatGPT and other LLMs as the gospel truth, rather than challenging their outputs.

Studies have shown that LLMs like ChatGPT can exhibit group biases similar to those observed in humans.

Research from Nature Computational Science reveals that AI models, including those used in language systems like ChatGPT, are prone to a well-documented human behavior known as “social identity bias.”

“Artificial Intelligence systems like ChatGPT can develop ‘us versus them’ biases similar to humans – showing favoritism toward their perceived ‘ingroup’ while expressing negativity toward ‘outgroups,’” explained Steve Rathje, a postdoctoral researcher at New York University and co-author of the study.

Furthermore, AI language models have been known to perpetuate “covert racism.”

They are more likely to suggest that people with African American speech patterns be assigned lesser jobs, convicted of crimes, or even sentenced to death.

Unlock more exclusive Cybernews content on YouTube.