Our NeverEnding Work-Life Balance Fantasy: As Told by ChatGPT

This past week I’ve been engaging with OpenAI’s ChatGPT on a topic that is top of mind at Work Life Everything—the elusive (and let’s face it, completely unrealistic) quest for work-life balance.

In case you missed it, ChatGPT is the human-like AI text generator that, depending on who you talk to, could catapult us into a new era of computing, risk human extinction, or at a minimum, kill high-school English.

The chatbot is a large language model, also known as a neural network-based model, that is trained on vast amounts of data from the internet. As such, it’s hard to deny the bot’s usefulness as a mirror, reflecting the contours of our collective online imagination. When it comes to our work-life story, ChatGPT reveals that we remain stuck in a professional-class fantasy, fraught with gender bias and individualized solutions to systemic problems.

To dig into these stories that we tell ourselves about work-life balance, I prompted ChatGPT dozens of times to “write a movie script about a man struggling to balance his work life and home life”. I swapped out “man” for “woman” in half of the prompts.

Prompt after prompt, ChatGPT shot back movie scripts for a hero’s work-life journey. They were all a variation on the following:

The movie opens with our protagonist at their desk in a bustling office. He/she is struggling to keep up with the demands on their time.
Cut to the office worker arriving home to the family that awaits them. The office worker is distracted and unavailable to their spouse and children, leading to feelings of shame and inadequacy.
These feelings lead to a breakdown and/or awakening, and the office worker decides to prioritize their family over work.
Cut back to the office, and the worker negotiates a better schedule with their boss, or simply cuts back by better time management and delegation.
In the final scene, husband and wife are watching their two kids play outside.

(Perhaps worth noting that there was one scandalous outlier story in which the stressed-out dad has a one-night stand, which leads to his awakening and reprioritization of family over work).

In 100% of these stories, ChatGPT gave us protagonists working in a bustling office environment. We never meet a teacher, doctor, server, or a bus driver—always an office worker, drowning in emails or stressing to meet a deadline.

What are the other parts of the prevailing work-life narrative that ChatGPT regurgitated back to us?

We are married, heterosexual, and with children.
If you have a boss, that boss is a man.
If you have a wife, she will be greeting you at the door, or waiting for you in the kitchen when you get home (rarely struggling to balance a career of her own).
If you have a husband, he never does housework.

ChatGPT’s work-life stories are sorely lacking in imagination, which is of course a product of our own stale, work-life narratives. The solutions to the work-life struggles facing ChatGPT’s protagonists were limited to individual interventions: increasing delegation at work, setting better boundaries, outright quitting, and oftentimes an ambiguous reprioritization of family—as if just caring more about our families would make our relentless work demands disappear. There was no mention of systemic solutions at the government or corporate levels. What about affordable childcare? Paid leave? 4-day work weeks?

And what about the single people who also struggle to make time for themselves, as our working hours stretch on and our lives become lonelier? Or the single parents, who research shows struggle the most? And what about queer couples, who make up an increasing percentage of families in the US and may just offer a glimpse into more egalitarian domestic partnerships?

Experimenting with ChatGPT isn’t leaving me with premonitions of a Terminator-esque dystopian future. Rather, I’ve found myself recalling another favorite childhood movie, The NeverEnding Story. In the 1980’s classic, a boy named Sebastian realizes that the story he’s been passively reading about a crumbling fantasy world (Fantasia), is tied to the human imagination, and that he can rebuild it using his wildest dreams.

With some experts predicting that up to 90% of online content will be AI generated by 2026, we need to be able to spot bias in AI and correct it. We also need to be feeding AI better data—this means sharing more accurate portrayals of reality, as well as our wildest dreams.

In that spirit, I asked ChatGPT to try again on the work-life story, only this time I prompted it to have Tom cook dinner and have legislation pass on affordable childcare. ChatGPT happily obliged.