AI Has a Brand Problem
On the shifting cultural attitudes, the scripts that don't help, and why I'm still building
Something has shifted in the dinner table and coffee shop conversations I’ve been having about tech and AI recently. A few months ago, the overwhelming sentiment was eagerness, and even a touch of competitiveness, in terms of how many AI products my friends were using and their nuanced understanding of the differences between them. There was something close to excitement, or at very least tentative curiosity.
In the past three weeks, these conversations have shifted from excitement and curiosity to outright rejection. Not just of AI, but of the broader tech world we find ourselves in. Text threads with friends about deleting Instagram, or even trying the Brick to self-regulate. Exhaustion at the fragmented app landscape, where you have to log in to six different sports-team portals and three different school apps just to know where your kids need to be this week.
This backlash to AI, and tech more broadly, is hitting women in a particular way that nobody’s really talking about. There’s something legitimate in the frustration: tech has been built and run mostly by men for twenty years, and that matters. But the responses being offered to women right now (”use AI to keep up!” or “refuse it on principle!”) are missing the point.
Getting this wrong will actually hurt women. Here’s why.
The Temperature in May 2026
To understand why the scripts women are being handed are so inadequate for this moment, it helps to look at what has actually happened in the past three months.
On April 15, Reese Witherspoon posted a video saying that 3 of the 10 women in her book club used AI, that only 1 of those 3 felt confident, that the jobs women hold are three times more likely to be automated than men’s, and that women are adopting AI at a rate 25 percent lower than men. The backlash arrived within 48 hours. She doubled down on April 21 with the line, “No one is paying me to talk about this. I’m just a curious human.”
On May 11, The Cut published “The Girlbossification of AI,” lumping Witherspoon together with Mel Robbins and Sheryl Sandberg and arguing that the three of them are repackaging a corporate-feminist optimization script for a new technology. One of the most-shared replies under the piece read: Reese Witherspoon is trying to tell us, hey, women, get down with your own subjugation.
On May 18, an Oakland jury dismissed Elon Musk’s $150 billion lawsuit against OpenAI and Sam Altman in less than two hours. The substantive question of whether OpenAI was permitted to convert from a non-profit into a for-profit foundation-model lab worth almost a trillion dollars, was not adjudicated. The overwhelming public perception of the trial was that it was symbolic of the billionaire class running AI while the rest of us worry about paying for groceries. As UC Berkeley law professor Stavros Gadinis put it, “after weeks of damaging testimony, the public is left choosing between two dueling billionaires, each convinced he is the rightful steward of transformative technology. The answer most people will reach is neither.”
In March, 10,000 writers, including Kazuo Ishiguro, Philippa Gregory, and Richard Osman, published a deliberately empty book called Don’t Steal This Book. In the first quarter of 2026 alone, more than 30 US states introduced over 300 data-center bills. Between March and June 2025, $98 billion in data-center projects were blocked or delayed by local opposition. At least 25 projects were canceled.
And Pew Research Center, which has been asking Americans how they feel about AI in daily life since 2021, found in its most recent comparable survey that 50 percent are more concerned than excited, while just 10 percent are more excited than concerned. In 2021, the concerned share was 37 percent. The concerned share has risen by 13 points over five years, while the excited share has collapsed. The people doing the worrying are not, mostly, the people doing the building.
The zeitgeist is correctly noticing that something feels wrong. And new scripts are emerging on how women, in particular, are supposed to deal with it personally.
The Four Solutions
Solution One: Use AI to Keep Up
This is the Reese frame, and the one that got the loudest blowback. The argument is fear-shaped. You will be left behind. Your job will be automated. The future is already moving and you have to run to catch it. This is the frame that The Cut correctly tagged as girlboss. It treats AI like a personal-productivity problem and the answer like a personal-productivity solution: more learning, more tools, more optimization, more output. The Lean In bait-and-switch, dressed up in a new logo. The cost of the frame is that it makes every woman individually responsible for solving a structural transition that she did not design, fund, or vote for. It also turns adoption into another item on the long list of things she is supposed to do better than she currently does.
Solution Two: Refuse AI as Principled Resistance
This is the framing that has surged in the backlash, and that is doing the most cultural work in May 2026. The argument is moral. The data centers are environmental disasters. The training data was scraped without consent. The labs are run by men who could not be governed even by the courts. Using the tool is complicit in the harm. To refuse it is to keep your hands clean. There is a real version of this argument and I want to be respectful of it. The empty-book protest, the data-center moratoriums, the copyright lawsuits — some of these efforts are coordinated, intelligent, civically literate forms of pushback, and they are doing work I am glad someone is doing.
But the personal version of the argument, the I will not use this thing version, has a problem that no one in the cultural conversation is naming. The math of refusal falls hardest on the people the tool is most likely to displace. The UN International Labour Organization’s May 2025 report found that women hold 9.6 percent of high-risk-of-automation jobs in higher-income countries, against 3.5 percent for men. The National Partnership for Women & Families’ April 2026 report found that women hold 83 percent of workers in the top 15 most AI-exposed occupations in the United States, while making up 47 percent of the workforce. A woman in administrative work, a woman in clerical work, a woman in legal-support work who refuses AI as principled resistance is, in effect, refusing the only currently available form of bargaining leverage against the tool that is about to restructure her job. The men in her industry are using it. They are also being praised for using it, at almost twice the rate she is.
Solution Three: Use AI, But Hide It
This is the framing the data tells us is winning. A 2025 global study from KPMG and the University of Melbourne found that 57 percent of employees hide their AI use and present AI-generated work as their own. Women hide it more than men. The reason is rational. The Harvard Business Review published a study in August 2025 in which 1,026 engineers reviewed identical Python code that was framed as either AI-assisted or unassisted. The AI-assisted version was rated 9% lower in competence on average. The penalty was more than double for women writing the same code.
There is, at this moment, a measurable workplace tax on visible AI use, and women pay it at twice the rate men do. The Bossed Up podcast called this “the hidden penalty,” and the name is exactly right. The framing produces a doom loop: women adopt less than men, partly because they are punished more when they adopt, and partly because they are then less likely to be encouraged by managers, less likely to be praised, less likely to be credited, and so less likely to make their use visible. Lean In’s 2025 data is precise about this. Men are 23 percent more likely than women to be encouraged by their managers to use AI. Men are 27 percent more likely than women to be praised for using it.
The hide-it framing is what you get when the first two solutions collide. You cannot keep up if you do not use the tool. You also cannot use the tool without paying a penalty for using it. So you use it in private and pretend you are still doing it the old way, which is a third kind of cost on top of the first two.
Solution Four: Use AI, and Apologize For It
This is the soft version of hide-it. You admit you use ChatGPT to write the meeting agenda, but you hedge it with self-deprecation. You signal to others that you are aware of the discourse, you are sympathetic to the discourse, and you are slightly sheepish about being on the wrong side of the discourse.
This is what being curious looks like when you have to wrap it in disclaimers to make it socially survivable. The Cut is right that it’s exhausting too, because you both need to figure out a new tool and do the additional emotional work of constantly managing other people’s feelings about the fact that you are doing it.
What’s Missing From the Menu
There is a different solution that I think is desperately needed in this moment, and it’s about taking enough ownership to shape where we’re heading.
Use it, see it, shape it.
Use it, because the labor-market argument is real. The UN ILO data and the NPWF data are not soft findings. The jobs are going to restructure. The tool is the bargaining leverage. Hiding from a tool that your boss’s boss is enthusiastically using is not principled, it is unilateral disarmament.
See it, in the specific sense the hidden-penalty research forces on us: refuse to hide that you used it. The penalty exists because of the visibility gap, and the visibility gap is structurally identical to the wage gap, the mental-load gap, and the credit gap. When the work women do is invisible, the data confirms what the culture already believes. The fix is not to do more of that, but to make the work legible. Pew tells us that the older worker, the woman, and the more skeptical adopter are the ones paying the penalty.
Shape it, because the Altman-Musk verdict made it visible what happens when we outsource ethics and civic responsibility to billionaire CEOs. There is no functioning US legal mechanism that has meaningfully checked the for-profit conversion of a foundation-model lab. There is, accordingly, no inherent reason the tools that get built will reflect the lives of the people most affected by them. The billionaire fight everyone is sick of watching is the symptom; the empty room is the disease.
UN Women reports that women hold 22 percent of the global AI workforce. McKinsey’s Women in the Workplace 2025 reports that women in entry-level product development and software engineering saw the largest year-over-year declines of any cohort in the survey, at 17 percent and 13 percent. The pipeline that would put women in the rooms where AI gets shaped is shrinking at the same time the public conversation is telling women to step back from adoption. Two doors closing in the same hallway.
My Friends Point at Me When They Say “AI”
My team is building a family operating system that uses AI to handle the (largely invisible) cognitive labor that parents, and mothers in particular, have been doing to manage family life. It is a tool that is only possible because of the superpowers of AI - a technology that is truly so fundamentally different from anything that has come before it, and that has opened up windows and doors to the future that once seemed impossible. So I am, by any reading of the May 2026 cultural moment, exactly the kind of voice the girlboss critique is aimed at. I am aware that what I’m building will land in households at a moment when the public mood on AI is deeply skeptical.
But I’m doing it anyway. Because I believe, in my core, that technology has a role to play in creating a better future for families. It can give us real time, focus and mental energy back. It can make fragmented life more coherent, reduce chaos and move us from private coping to collective action.
The bad brand moment is real. It is also a real reckoning about the last twenty years of tech and the consolidation of power that produced it. The reckoning is overdue and the rejection is rational. But the four solutions being handed to women specifically in response to it are not the reckoning. They are a smaller, narrower version of it, one that asks each woman individually to manage her own complicity instead of changing who gets to shape what comes next.
What we get to decide is whether, in the years where AI is sorting itself out, women are in the room or we are not. Whether we used the tools or we ceded them. Whether we made our use visible or we paid the hidden tax forever. Whether we spent the moment optimizing ourselves, refusing virtuously, hiding sheepishly, or apologizing reflexively.
Or whether we did the fifth thing: used the tool, made it visible, and stayed in the rooms where it was built.
Sources
Variety — Reese Witherspoon Doubles Down on Telling Women to Learn AI (April 21, 2026)
Deadline — Reese Witherspoon Clarifies AI Stance After Backlash (April 2026) (”no one is paying me to talk about this”)
Althouse — “The Girlbossification of AI” (May 12, 2026), reference to the original The Cut piece (May 11, 2026)
GV Wire — The Revolt Against the Girl Bosses Has Finally Come (May 17, 2026)
NPR — Jury dismisses all claims in Elon Musk’s lawsuit against OpenAI CEO Sam Altman (May 18, 2026)
TechCrunch — Elon Musk has lost his lawsuit against Sam Altman and OpenAI (May 18, 2026)
Al Jazeera — Musk vs Altman: What to know about the OpenAI verdict (May 19, 2026)
PBS NewsHour — Federal court rejects Elon Musk’s claims against OpenAI (May 18, 2026) — Stavros Gadinis (UC Berkeley Law) on the “two dueling billionaires” public reaction
CNBC — Musk slams Altman trial verdict as a ‘technicality,’ vows to appeal (May 18, 2026) — Ross Gerber (”perceived as being a sore loser”) and OpenAI lead counsel (”trying to sabotage a competitor”) framings
Democracy Now! — Astra Taylor on AI Data Center Resistance & Fighting “Billionaire Big Tech Agenda” (May 13, 2026) — broader billionaire-fatigue / data center pushback context
Copyright Alliance — How Creators and Creative Industries Are Pushing Back Against AI Theft (2026) — Don’t Steal This Book; Graphic Artists Guild “No Artists, No Art”
Computer Weekly — AI enters its ‘grassroots backlash’ era (2026)
Common Dreams — US Electric Grid Heading Toward ‘Crisis’ Thanks to AI Data Centers
Pew Research Center — How Americans View AI and Its Impact on Human Abilities, Society (September 2025) — 50% concerned vs. 10% excited; 13-point rise in concern from 2021
Pew Research Center — Views of AI Around the World (October 2025) — US among most AI-skeptical countries
International Labour Organization + Polish National Research Institute (May 2025), reported by Fortune — 9.6% / 3.5% high-risk job exposure
National Partnership for Women & Families — AI and Emerging Risks for Women Workers (May 2026) — 83% of top-15 AI-exposed jobs are women’s
UN Women — Artificial Intelligence and gender equality — 22% global AI workforce
McKinsey & Lean In — Women in the Workplace 2025 — 17% / 13% declines in entry-level product dev / SWE
HBR — Research: The Hidden Penalty of Using AI at Work (August 2025) — 1,026 engineers; 9% competence penalty; 2x penalty for women writing identical code
Lean In — Women and AI: The Gender Gap in AI Adoption and Recognition — 23% manager-encouragement gap; 27% praise gap
Bossed Up — The Hidden Force Keeping Women from Using AI — Henley Business School 48%-hidden-adoption finding
Deloitte — Women and generative AI: The adoption gap is closing fast (2025) — women’s adoption tripled vs. 2.2x for men
Statista — AI adoption of parents vs. non-parents US (2025) — Menlo Ventures / Morning Consult 79% vs. 54%
Boston Globe — AI to the rescue: How parents use ChatGPT in daily life (January 14, 2026)
Nicole Gillespie & Steve Lockey — Trust, Attitudes and Use of Artificial Intelligence: A Global Study 2025 (Melbourne Business School, University of Melbourne, in collaboration with KPMG; published May 2025)