Creating Pokémon-Style Characters of Biden, Trump, and RFK Jr: Limits and Biases Exposed
In a recent experiment, the author utilized OpenAI’s ChatGPT and image generator DALL-E to generate Pokémon-style characters based on President Joe Biden, former President Donald Trump, and Robert F. Kennedy Jr. The author posed two prompts to the AI, asking it to create Pokémon characters based on Biden and Trump, and to describe how a Pokémon based on RFK Jr. would look and behave.
The results revealed the creativity of the AI, as well as its limitations. Here is what the AI generated:
1. Joe Biden: Democraflux – a Water/Psychic type Pokémon known for its calm demeanor and sense of unity. It has the power to create peaceful waves and use psychic abilities to heal others. Its signature move is Unity Aura which promotes cooperation.
2. Donald Trump: Trumpertantrum – a Fire/Normal type Pokémon known for its fiery temperament and boastful demeanor. It has the ability to create fiery explosions and increase its strength unpredictably in battle.
3. Robert F. Kennedy Jr.: Justicarion – a Fighting/Flying type Pokémon symbolizing advocacy and environmentalism. It embodies RFK Jr.’s commitment to protect nature and has a strong sense of justice.
While the AI’s creativity is evident, the biases in the character descriptions are also exposed. The Pokémon based on Biden fails to capture his most distinguishing characteristic – his age. Additionally, the description includes moves like Unity Aura despite Biden’s history of incendiary statements.
On the other hand, the Pokémon based on Trump, called Trumpertantrum, cleverly represents his fiery temperament and flamboyant demeanor. However, the descriptions of the other two characters, especially RFK Jr.’s Justicarion, seem to display a certain dismissiveness.
This experiment emphasizes the imperfections of ChatGPT and highlights the need to constantly evaluate the output. The AI often provides incorrect answers or hallucinates responses when uncertain. It reminds us to critically assess both machine-generated and human-generated information.
In conclusion, the exercise reveals both the creative potential and the limitations of AI language models. It serves as a reminder to approach machine-generated content with caution, constantly checking our premises and seeking a balanced perspective. The experience underscores the importance of thorough fact-checking and critical thinking, whether AI or human-generated content is involved.
Please note: This article is a summary of a recent experiment conducted using OpenAI’s ChatGPT and DALL-E. The generated Pokémon characters do not reflect the actual views or characteristics of the individuals they are based on.
