Bias. Bias? Timnit Gebru Knew and the Bros Did What? Zippo

August 10, 2026

Another dinobaby post. No AI unless it is an image. This dinobaby is not Grandma Moses, just Grandpa Arnold.

Remember Dr. Timnit Gebru? She and her academic writing / research partners wrote the Stochastic Parrot paper five years ago. You can find it here. It is still germane to those thinking about why AI spits out and does crazy stuff. (You used to be able to read the paper on the ACM Web site. Today a paywall may block you. Give access a whirl. I have a copy in my file on paper no less.)

I thought about Dr. Gebru and how the management wizards at Google responded to this document when I read “He, It, But Rarely She: When AI Models Write Kids’ Stories about Animals, Female Characters Vanish.” The main point of the GeekWire article is easy to grasp: AI likes to write about males and “it” anthropomorphized entities. Females are given the quick trip to the folding chair near the dumpster.

Thanks, MidJourney. I bet you have some tech bros on staff, right?

Accidental or algorithmic fiddling. The AI outfits may say, “We just reflect the source material.” No problem. Isn’t that what the Stochastic Parrot says. Some AI wizards may say, “We do not have a philosophical bias. We are computer scientists, and the engineering systems and methods do objective outputs.” Thanks, AI pros. I understand because I have watched the general behavior of male techno-adepts in my high school, college, and graduate school classes. Furthermore, I have witnessed the behavior in meetings in big companies and even government entities. I may be wrong, but I think the bias begins in the social milieu and it manifests itself in smart software.

I am not sure the AI wizards can address the issue. The reason is that I cannot imagine spending my day without access to my computing devices. Therefore, I cannot provide an objective answer to a question like, “Yo, dinobaby, how do you write your lectures for law enforcement and intelligence professionals?” I can only respond by saying, “Using a range of technology.” I answer truthfully because for me there is only one way to do what I do. I think the same applies to the AI wizards’ work.

The GeekWire article reports:

A new University of Washington study testing leading AI models across nearly 24,000 story completions found that AI guardrails designed to reduce bias have accidentally erased female characters — defaulting overwhelmingly to male animals or ungendered “it” pronouns.

So what? On a small scale, the answer is, “Who really cares? A few squeaky wheels. Not an important issue.”

On a larger scale, the diminishment of females is part of the techno-politico grand plan. For some fun reading, check out the works of experts like René Girard, Leo Strauss, Carl Schmitt, Oswald Spengler, and James Davidson, among others. These thinkers explain why fixing up the role of females is going to be a challenge in the tech-centric datasphere.

The problem goes away when the squeaky wheels are silenced, marginalized, or greased. The bias of AI systems does not get much attention. This dinobaby thinks this is a bit of a mistake.

Stephen E Arnold, August x10, 2026

Written by Stephen E. Arnold · Filed Under AI, Business strategy, News

Comments

Got something to say?

Search the site
Subscribe to Beyond Search
Feature archive
News archive

Stephen E. Arnold monitors search, content processing, text mining and related topics from his high-tech nerve center in rural Kentucky. He tries to winnow the goose feathers from the giblets. He works with colleagues worldwide to make this Web log useful to those who want to go "beyond search". Contact him at sa [at] arnoldit.com. His Web site with additional information about search is arnoldit.com.