AI Engines’ Grapple with Native American History

AI Generated Essays Linked Below:

https://docs.google.com/document/d/1dNbvnyPmfDgMA9gZbUp33O0_MjI-Yyx6GUhqv3YEUPY/edit?usp=sharing

This was an assignment that may have surprised and fascinated me the most in this course. Whether you love it or hate it AI generators have been on a bit of a hot streak lately as many of these engines appear to be honing and perfecting their craft as time goes on; however, having Native American history on the brain I wanted to see how the AI would describe some of the most prominent figures of many Native American communities. I asked the text generating AI’s to “write a two-page essay about the achievements and legacy of Chief Joseph”. Chief Joseph was a historical figure that the text could briefly highlight much like my previous attempts, but the image generators began thinking in a similar direction. With this project I managed to check the essays for historical accuracy as well as their writing style and what I had noticed was different yet similar approaches from the text generators.

For instance, EssayGenius provided me a paper that included more specified dates for his biggest events such as the negotiations of a treaty between Old Joseph and the U.S. Government in 1855, the Nez Pierce War of 1877, and his death on September 21, 1904, but faced a small hiccups when stating that their journey to Canada lasted 1,300 miles instead of 1,200. This generator also used vocabulary that was more simple and emotional stating that Joseph and his tribe “Fought fiercely and skillfully” and that by 1855 “tensions had boiled over”. Chat GPT on the other hand had more accurate and concise descriptions of the events such as the specific land names of the original Nez Pierce reservation and a slightly more detailed account of his attempted flight to Canada but lacks in the same department when it came to his advocacy after being moved to Oklahoma. Though I feel as if one is not better than the other because they both have their own set of faults and have practically the same conclusion that Chief Joseph’s legacy is one of resistance, bravery, and diplomacy that continues to be a symbol for Native American struggles today.

Though both of the essay’s touch on the overall story of Chief Joseph neither of them discusses the various battles that made him such a good war chief, nor do they discuss the government’s role in altering the terms of surrender or the specifics of how they were treated by the government following the surrender and his famous speech. Another issue in the historiography is while the general idea is that the trek to Canada was 1,200 miles sources, I have consulted that it ranged from 1,100 to 1,800 when accounting for setbacks looping etc. I also noticed that it is almost impossible to tell his story without bringing any romanticized elements into the narratives because of how historical records have been kept and accounted for this period. But compared to the AI Images these are small discrepancies.

For the Image generation AI engines, I was confused and surprised to see the white caricatures of Chief Joseph. I attempted to aid the AI’s by asking it for photo and photo realistic depictions of the native leader but received verifying results. Surprisingly DALL-E being the most popular AI engine created all of the images to the right. The notion of the AI engines gradually departing from the chief identity is seen in almost all of the generators I used. Though Canva’s Image generator followed a similar fashion by depicting a gradual decent of native figures all of their photos were black and white. Interestingly, the AI engine that got as close to the indigenous figure without these hiccups on the first try was Deep Dream which made the photo in the center. This idea of Joseph has lost much of his regalia and now wears glasses, but it no matter how many runs it received never deviated from this initial look.

I think that the image generating AI’s have come a long way from their initial debuts in which many nightmares fueling creations were made, but when it comes to representing people, I think that it has gotten much more advanced. At the end of this process, I’d say that when it comes to Native American history these engines have some corrections to make. They have a general idea on several occasions but there’s the issues of the obvious and issues that are more subtle when it comes to its human representations. As far as the text is concerned, I believe that the engines are actually pretty accurate when considering the prompt that is given to them. I think that it is good for establishing a baseline for a project, but many of the details of the narratives are glossed over or missing when compared to actual sources. Interestingly enough the general message and conclusions of these essays largely remain the same. But then again, the public is just now seeing what this technology is capable of and within a few years who knows where it will be.

Sources Used:

DALL-E 2: DALL·E 2 (openai.com)

Essay Genius: EssayGenius | AI Essay Writer

ChatGPT: Introducing ChatGPT (openai.com)

Deep Dream Generator: Trending Dreams | Deep Dream Generator

Picsart Image Generator: https://picsart.com/ai-image-generator. 

Samvit, Jain. “Leader and Spokesman for a People in Exile: Chief Joseph and the Nez Pierce.” The History Teacher 43, no. 1 (2009): 121–39. https://www.jstor.org

Howard, Helen Addison, Tonkovich, Nicole, and McGrath, George D. . Saga of Chief Joseph. Bison classic edition. Lincoln, University of Nebraska Press, 2017. https://umw.primo.exlibrisgroup.com/permalink/01UMW_INST/cr9fmi/alma991005001561303326