If there are any universities left in America that prioritize the education of students and devote themselves to maintaining traditional family values, Brigham Young University ought to be one of them. While I still believe in BYU, it is disturbing that, even here, many of my class discussions operate on the assumption that raising families is a side gig or even a burden that distracts from your “true mission” in life.
Recently, in a course on machine learning, our professor addressed the topic of “gender bias” often exhibited by word embeddings, an important component of artificial intelligence language modeling systems like Google Translate and ChatGPT. Put simply, a word embedding roughly models various aspects of a word’s meaning and usage with a vector of numbers. We can add and subtract word embeddings to and from each other to reveal the relationships the language model learned from the words in the dataset. An oft-used example involves the words king, man, and woman. We can take the embedding KING, subtract MAN, and then add WOMAN:
KING – MAN + WOMAN = QUEEN
In a successful language model, we typically expect the resulting embedding to be very close to the one the machine created for QUEEN. In short, we want our model to have learned that man is to king as woman is to queen.
Our professor drew our attention to a 2016 study where researchers discovered interesting associations in language models trained on datasets compiled from Google News articles. Performing a similar experiment as above, they swapped king with computer programmer:
COMPUTER PROGRAMMER – MAN + WOMAN = HOMEMAKER
Thus, based on patterns tracked in Google News articles, a model deduced that man is to computer programer as woman is to homemaker.
Typical to my experience in academia, our professor led a congregational lamentation of the implicit sexism and “deep problems” in our culture that such a study reveals. Many classmates bemoaned that we as a society believe that women cannot fill occupations typically dominated by men, and should be compelled to be stay-at-home moms.
I was astounded that these students truly believed that modern Western society states women cannot or should not do things like program a computer; I would dare any of them to take a poll on the street and prove this claim. I don’t think I’m too presumptuous when I say that women need fewer guys championing for them to be behind a computer, and more gentlemen to get out and meet them. However, two things truly trouble me.
First, I struggled to understand why the class seemed to think that there lies some glory or honor in the headbanging, frustrating task of computer programming. Unlike most group members, I have been a working professional for nearly two years. Sure, this isn’t a long time, but it’s been enough for me to have learned something most university students don’t understand: most jobs are only jobs. Sometimes, you’re not changing the world or “living a dream,” but you’re doing something that simply needs to be done. Work, for most people, is just work. I don’t say this to demean my work or anyone else’s work, and I don’t mean to be pessimistic; our society is built on the little contributions we all give daily. But, I think we need to be careful about giving young people the idea that their careers will be magical and fulfilling. It takes some humility to see that, for most of us, the meaning we truly seek in life will likely be much closer to home. This brings me to the second and greater point:
Why did the class accept the premise that being a homemaker and mother is of less value than a career in computer science? At a religious institution like BYU, I imagine if I were to ask that question point blank, most would deny it and proceed to praise their mothers. But, if they genuinely believed motherhood and homemaking are of great value, why would they get upset when a machine finds a connection between womanhood and homemaking? Indeed, a woman is free to pursue education and career as she chooses; however, I think it’s wrong to assume that a woman’s decision to be a mother and homemaker is merely a result of oppression. Such a belief presumes that raising children is inglorious, burdensome, or an obstacle in your true path in life, demeaning women everywhere who willingly sacrifice a career for their children. The world may say that AI systems are biased and our culture is sexist, but I argue that it is the nature of humans to want to raise offspring.
For previous generations, raising a family was assumed to be the primary purpose that both men and women would fulfill. A woman’s nurturing of children was noble; a man’s working to provide for and protect a family was noble. That outlook, however, is disappearing. I blame this in no small part on the universities, BYU included, who plant visions in their students’ heads of changing the world through activism and careers and then send them to chase fulfillment in the wrong places. But men are meant to be fathers just as women are meant to be mothers, and we need each other. No other roles are of greater worth.
This should be no surprise to disciples of Jesus Christ who know the divine and central role of the family in God’s plan; indeed, we devalue motherhood and fatherhood at our great peril. There is something exquisite in the relationship between parent and child; it certainly provides a far higher purpose than building things with machines. And if a machine finds patterns in our data that reflect traditional family ideals, that is not a “deep problem.” That is something good, even beautiful. The restored Church of Jesus Christ has taught the divine and central role of families in God’s plan since its inception, and I hope at BYU we can learn to dispel the cultural influences that teach otherwise.
Written by: Lon Hedgerman
The Cougar Chronicle is an independent student-run newspaper and is not affiliated with Brigham Young University or The Church of Jesus Christ of Latter-day Saints.