My husband and I often disagree about the mental load. From his POV, he’s a very participatory parent who does a lot of the heavy lifting. He handles the cooking, the finances and the majority of the kid shuttling—speech therapy on Tuesdays, swimming on Saturdays. But the friction comes down to a teeny-tiny distinction: I’m the one doing most of the remembering. ( “What are we having for dinner tonight again?” “Did we pay that pediatrician bill?”)
To his credit—and mine—we are both busy. We each work full-time and have two kids at wildly different ages and stages (my third grader has 45 minutes of homework a night; my two-year-old has a weekend birthday party rotation that just won’t quit). But I’ve always been better at the details, so those have tended to fall to me, resulting in a system that doesn’t feel fair.
After spending years talking to my therapist, my mom friends and experts on the topic (hi, Eve Rodsky), I decided to challenge a truly impartial judge to dig into the data: ChatGPT. How, exactly, are my husband and I dividing the invisible household work? And could an unemotional audit help us find a better solution? What I uncovered unlocked a less-talked about truth in our partnership—and even better, a tactical solution.
I Ran an AI Audit on My Marriage and the Data Changed Everything
The robot has spoken
How Chat Entered the Chat
I’ll preface this by saying I’m skeptical of AI—I prefer hashing things out with an actual human. Always have, always will. But one area where AI really shines is data, and I was curious how it would feel to exhaustively spell out who does what. Then, using an analytical approach, see if we might glean a solution that, ultimately, helps us get time back.
I kept things broad with my prompt: “The mental load in my marriage is a challenge. Is there a way to audit the work we do and come up with a better system?”
Chat kicked things off with an exercise to capture everything with a lengthy list of questions about who does what—ranging from grocery shopping to managing holiday traditions—with the goal of figuring out, “who is the default household manager across tasks,” and, “are shared tasks really shared or is one person coordinating everything”? It also aimed “to uncover responsibilities that could be eliminated, automated, outsourced or simplified.”
After answering the prompts, we saw that it was almost a 50-50 split. But as it turned out, not all work was created equal.
ChatGPT Then Analyzed Our Contributions, Category by Category
For example, household basics (things like cooking, laundry, general tidying) illuminated a mix of shared responsibilities, which surprised me: I had initially credited myself with doing most of the work.
But the next category, family/kids (everything from buying gifts to those aforementioned birthday parties to managing pediatrician appointments) was firmly in my court. The AI agent even interjected here to highlight a distinction: “Your husband has meaningful executive ownership, but you appear to be carrying the executive function layer.” I.e.: He was available to discuss kid-things, but I was the one prompting the conversation.
ChatGPT blew my mind with its summation: “This issue doesn’t sound like, ‘My husband doesn’t do enough’; it sounds more like you have become your household’s anticipatory system.”
Next up, was money/admin/household management (taxes, vacation-planning, car-stuff), then finally social, family and emotional logistics (maintaining friendships, remembering family traditions, planning date nights). My husband predominantly carried the first, I managed the second. But I noticed that I was writing “both” across several to-do’s, which is where I got called out. “This is probably the most important finding,” the robot offered. “You have several things marked ‘both,’ but the coordination still sits with you.”
ChatGPT then blew my mind with its summation: “This issue doesn’t sound like, ‘My husband doesn’t do enough’; it sounds more like you have become your household’s anticipatory system.” Oof.
This Final Gut Check Was Truly Insightful
After parsing through all the data, ChatGPT said it was devising a new and improved household system for us to try out. But it wanted the answer to a very telling question before it could start: “What thing about the current arrangement makes you most resentful, exhausted or trapped?”
My answer surprised even me: I’m burnt out on constantly having to scaffold my spouse.
Take swimming. My husband is the one taking our son, finding nearby parking and ensuring that he gets there on time. But I’m often the one packing the swim stuff, reminding everyone to have a snack beforehand, even just prompting my spouse to confirm the lesson is still on.
ChatGPT offered a “eureka” moment here: I don’t need my husband to do more. I need him to fully take ownership.
There *Is* a Solution
At one point in my marriage—and before children—I adopted the approach that we should each play to our strengths. For example, my husband is a way better chef than I am, which means he should handle the cooking. But I excel at family organization. Therefore, household management should fall to me. The problem is that household management is too much of a catch-all. And with the addition of small children, it becomes immeasurable.
By making myself the default family member for household infrastructure, I had taken control in a way that left little room for outside problem-solving. The backup plan was always me.
And AI had a gentle criticism of my efforts: Without full and clear ownership of tasks, I had adopted a prevention strategy that was also taking up mental space. In other words, I was working overtime to make my husband’s unpredictability—say, triple checking that we had all the ingredients for dinner—predictable.
That had exponentially increased my mental load.
As a solution, it recommended that we come together to review the audit, designate clear ownership, but most importantly, externalize our systems with what AI called “reliability guardrails”.
For example, my husband owns the meal plan. That means that he has to lay out four to five dinners by Sunday night. The meals all have to meet a weeknight feasibility rule of 30 to 40 minutes and he needs to generate a grocery list at the outset and provide me with that list. If there’s a missing ingredient day of, he handles it. But he also needs a built-in back-up plan. Say, a frozen pizza or even takeout.
The goal: I’m never in the position of becoming “an emergency operations center for family dinner at 5:30 p.m.” He owns it full stop and our externalized system makes room for error without the error becoming my problem.
What I Actually Learned From My Robot Therapist
ChatGPT was clear: When someone owns a task, it’s not how it gets done, it’s that it gets done.
In other words, I have to let go. That means accepting some inefficiencies without commentary. If he owns dinner, I can’t simultaneously own dinner by quietly checking the grocery list, reminding him about the meal plan and swooping in when something goes sideways.
I appreciated the final line in our chat, which sums up my thoughts exactly: “This is not an indictment of your husband. He’s contributing substantially and genuinely overloaded, too. The problem is that your current system makes your executive functioning compensate for his, and that arrangement is costing you both.”
The funny thing is, I went into this experiment expecting AI to tell us how to divide and conquer more fairly. Instead, it confirmed that fairness wasn’t really our problem.
My husband was already doing a lot. I was, too. The problem was the distinction between participation and ownership. He was participating in plenty of the work, but I was still carrying the cognitive responsibility. And because I’m good at anticipating problems, I had slowly turned that into my job—and my job only.
That might be the biggest lesson here: If I want my husband to truly own something, I have to give him the space to execute, even if execution isn’t always perfect. And if he’s going to own it, I have to stop quietly owning it, too.
