0:11
Standing Out From the Single-Stock ETF Crowd
As the Architect in The Matrix Reloaded put it: “the problem is choice.”
That’s particularly the case nowadays for leveraged and inverse single-stock funds, which are coming to market en masse to take advantage of some of the hottest companies on the market. Most recently, a flurry of single-stock ETFs have launched offering exposure to SpaceX, but similar treatments have been given to NVIDIA, Tesla, Apple, Coinbase and MicroStrategy. With Anthropic and OpenAI IPOs on the horizon, we can expect to see more soon. It’s making for a complicated and confusing menu of funds for advisors looking to add leverage.
“Marketing is huge in this space,” said Steve Foy, senior vice president of trading at Tidal Group. “Awareness is really the true differentiator.”
These types of funds have a short-term nature, so investors aren’t committing to an investment they’ll hold beyond a day or so. But they still have to decide between funds that are essentially offering the same thing, and often, for the same price. Morningstar analyst Zachary Evens agrees that brand awareness is key for issuers jostling for a spot in an investor’s portfolio.
“There are a handful of single-stock providers that are fairly popular in the retail community, and the assets reflect that,” he added. “Brand awareness would make it potentially likely for a trader to go to one brand first over another brand just because they’re aware of this brand over another that might offer the same or competing product.” Morningstar’s database shows that there are 430 single-stock ETFs in the US, with the largest providers being Direxion, GraniteShares and AXS Investments.
But branding isn’t the only differentiator:
- Speed to market is another major factor “for everything that’s not the biggest stock story of the day,” Foy said. “It’s about being there first and being available.” That could mean offering a single-stock fund for a stock that’s not yet hotly watched, though it also means focusing on new, in-demand areas of the market.
- Volume also matters, since investors want to be invested in a tight market. Tax management can play a role, too, Foy said, since investors typically don’t want to catch a dividend when only owning a fund for a few days.
Risky Business: Single-stock ETFs can generate high returns, especially when offering 200% (or more) of the performance of an underlying stock that’s taking off. But they come with plenty of risks, including the potential for outsized losses. Evens said to also be wary of volatility decay, a phenomenon where even if the underlying stock price goes up over time, the ETF doesn’t gain the same amount and might actually lose money. “That’s why these products should not be held for extended periods of time,” he added.
The post Standing Out From the Single-Stock ETF Crowd appeared first on The Daily Upside.
3:22
OpenAI, Anthropic Speed Toward IPOs Amid Growing Scrutiny of Token Payments
America’s leading AI labs are set to spend the second half of this year preparing for initial public offerings that will vault them into instant megacap status. Anthropic, valued at $965 billion in May, is slated to debut as early as October, and OpenAI, valued at $852 billion in March, is likely to follow in 2027.
As listings loom, questions about the nature of the companies’ token-payment business model are getting louder. Palantir CEO Alex Karp blasted the model last week, telling CNBC that “something has gone completely wrong.” Enterprise customers have begun to question the burdensome cost of pay-per-use token consumption and look to cheaper, less sophisticated open-weight models. So can the world’s two great AI labs keep up the pace?
The aggressive adoption of agentic AI in some workplaces this year led to the creation of the slang term tokenmaxxing. That’s what happens when engineers, under pressure to demonstrate they are integrating the new technology but with no clear guidelines, use AI models to excess. Executives have quickly realized this isn’t the most efficient way of doing business. As a result, companies including Uber, Microsoft, Salesforce and Meta have taken steps to ration their employees’ use of advanced AI because the token payment structure preferred by Anthropic and OpenAI has proven more expensive than it’s worth.
Speaking to TBPN, Palantir’s Karp said the excessive use of AI without regard for whether it creates value is “kind of like a porn addiction.” During his CNBC appearance, he said the US AI industry should not dismiss the potential for cheaper open-weight models, especially those in development in China, to close the gap:
- Beijing startup Z.ai’s GLM-5.2 model is now ranked among the top 10 large language models by Artificial Analysis, and is ranked as the second-best model for web development by AI evaluation platform Code Arena, placing it alongside Anthropic, OpenAI and Google. The open-weight model is also four to six times cheaper than frontier AI.
- Some US and international enterprise customers have already reported switching to cheaper Chinese models like DeepSeek and cutting back on payments to OpenAI and Anthropic.
Raising the Stakes: The Financial Times reported last week that OpenAI has held talks with the Trump administration about giving the US government a 5% stake. But the paper said its proposal hinges on other US AI labs, like Anthropic, agreeing to do the same. Experts, meanwhile, warn that recent export controls on advanced US AI models may accelerate the international adoption of models developed by Chinese firms. Two respected Silicon Valley financiers who have the president’s ear understand this: Former Trump advisor David Sacks and current Trump advisor Marc Andreessen have both noted GLM-5.2’s power in recent weeks.
The post OpenAI, Anthropic Speed Toward IPOs Amid Growing Scrutiny of Token Payments appeared first on The Daily Upside.
6:43
Tesla’s Delivery Rebound Fails to Crush Negative Narrative
The biggest energy crisis in history just delivered a much-needed opportunity for Tesla.
Last week, the electric vehicle maker said it had delivered 480,126 cars in the second quarter, a massive increase from just 384,122 deliveries a year ago and blowing out consensus analyst expectations of around just 406,000. On Wall Street, however, Tesla is stuck in reverse: Shares fell more than 7% on the same day as the big beat, marking its worst single-day slide in about a year.
Spinning Its Tires
Even with the big sales resurgence, Tesla likely won’t touch its annual deliveries peak in 2023, when it sold 1.8 million cars. Last year, its deliveries slumped to about 1.6 million; this year, it looks on pace to return to about 1.7 million. The causes of its sales slump are numerous and obvious: the death of the $7,500 EV tax credit in the US, rising competition abroad and the cancellation of its much anticipated low-cost model. Oh, and a lingering Musk-y smell of politically charged bad PR.
External estimates indicated that Tesla’s US sales remained in a slump this quarter, while surging in Europe and China, where the energy crisis has been even more acute. Still, it wasn’t enough to catch a key competitor:
- Cox Automotive said its data showed a 20% year-over-year decline in US deliveries in the second quarter. Meanwhile, new vehicle registration data in Europe showed Tesla sales rose around 57% in the region through the first five months of the year, according to ACEA, and sales in China have been on an eight-month growth streak, according to the China Passenger Car Association.
- Rival Chinese EV-maker BYD also reported its second-quarter deliveries last week, trouncing Tesla with 557,090 vehicles delivered over the same three-month period. Its overseas sales were up more than 94% in June as it speeds into Europe and Latin America.
Thrive to Survive: EV also-ran Rivian reported its deliveries last week, too, claiming 12,194 in the second quarter, above expectations. Take it as a sign that the EV winter may be thawing. Tesla, which saw investors “buy on the rumor, and sell on the news” of its big beat, is hoping the US launch last week of its Model Y L, which features a third row of seats, can help its turnaround. In other words, the way back for Tesla involves the revival of “the way back” seating option.
The post Tesla’s Delivery Rebound Fails to Crush Negative Narrative appeared first on The Daily Upside.
9:30
Strategy’s Days of Never Selling Bitcoin Are Done
Strategy founder Michael Saylor is famous for saying “never sell,” but as crypto craters, he has changed his tune. Bitcoin holding company Strategy is considering selling up to $1.25 billion in bitcoin after first parting with 32 tokens in late May.
The company is trying to win back investor confidence as cracks appear in its buy-buy-buy strategy. The so-called infinite money glitch that Strategy relied on, mainly issuing equity and buying bitcoin with the funds, has struggled to keep spinning under bitcoin’s prolonged downturn. Bitcoin briefly dipped below $60,000 last week, down from its October peak of more than $126,000.
Saylor’s backtracking is part of a wider plan to overhaul Strategy’s financing model.
Not Smooth Sayling
Strategy has amassed more than 847,000 bitcoin, buying more tokens even as bitcoin’s price climbed higher and higher. It has paid an average of $75,000 per token, about $15,000 more than the level where bitcoin’s price has been hovering. But the company continued to pay dividends to its shareholders, depleting its cash reserves. Now, it’s trying to build up cash to make sure it can meet its obligations even if bitcoin takes longer to bounce back:
- In addition to considering additional bitcoin sales, Strategy will buy back up to $1 billion of its preferred stock and up to another $1 billion of Class A shares. It’s also raising the dividend to 12% on its most popular preferred stock called Stretch. Strategy said it’ll bulk up its cash reserves to cover 12 months.
- The overall plan serves to boost the confidence of investors doubting whether Strategy can keep its commitments while bitcoin continues to reach new bottoms for the year. Strategy’s common shares climbed back above the key inflection point of $100 last week.
Two-Way Street: JPMorgan analysts raised a yellow flag on Strategy’s new plan, saying that bitcoin’s biggest corporate holder giving itself the option of selling as needed adds more uncertainty to a market already rife with FUD (a crypto-ism for fear, uncertainty and doubt). When Strategy buys, other investors feel emboldened to do the same. And when Strategy sells, well, some investors are likely to follow its lead then, too.
The post Strategy’s Days of Never Selling Bitcoin Are Done appeared first on The Daily Upside.
12:06
Comcast’s Spinoff Highlights a Growing Divide in Telecom ETFs
We’ll take Door No. 3.
The Comcast spinoff made waves last week, but it’s also creating new interest in exchange-traded funds that gain exposure to the telecommunications sector. The funds generally track companies like landline and mobile phone carriers, as well as mobile phone manufacturers, but AI companies are quickly splitting the segment in two. Behind Door No. 1 are traditional, mid-size telecom companies like Comcast and AT&T, which are struggling from good, old-fashioned competition. Behind Door No. 2 is the large-cap, new wave of companies like Meta, Alphabet and SpaceX, spending heavily on their AI buildout and often more volatile. It’s creating an interesting investing dynamic for advisors looking to leverage telecom.
“Meta and Alphabet are increasingly being connected to the AI trade, in or out of favor, whereas much of the rest of telecom services is slower growth,” said Todd Rosenbluth, Vettafi head of research.
In sector ETFs, fund performance largely comes down to portfolio composition. The State Street Communication Services Select Sector SPDR ETF (XLC), for instance, is down more than 6% this year, with about 40% of the fund allocated to Meta and Alphabet, both of which have slid amid concerns about AI spending. By contrast, the State Street SPDR S&P Telecom ETF (XTL), which holds neither Meta nor Alphabet and instead tracks more traditional telecom services, is up almost 48% over the same period.
The largest communications ETFs, according to Morningstar data, are:
- The State Street Communication Services Select Sector SPDR ETF (XLC), the dominant player by a wide margin, with nearly $22.8 billion in assets under management. XLC is down about 6% in 2026, as of the closing bell on Thursday, according to ETF.com.
- The Vanguard Communication Services Index Fund (VOX), which comes in second, with almost $5.7 billion AUM. VOX is down 2% this year.
- The Fidelity MSCI Communication Services Index ETF (FCOM) takes the third place, with $1.7 billion AUM. The fund is also down about 2%.
The good news is that XLC brought in $337 million in net flows last week after bleeding more than $1 billion in assets over the month of June, which Rosenbluth said may signal improved investor sentiment.
Press S for SpaceX: SpaceX used to be viewed as complementary to the sector rather than competitive. But the aerospace and rocket manufacturer recently signaled bigger ambitions to expand its broadband and wireless services from rural areas into more densely populated markets, bringing it into more direct competition with traditional telecom, said Michael Hodel, Morningstar equity director. “It’s not necessarily a good hedge to own SpaceX as a way of balancing the risk around traditional telecom stocks.”
The post Comcast’s Spinoff Highlights a Growing Divide in Telecom ETFs appeared first on The Daily Upside.
15:20
SEC’s Interest in Novel ETFs Could Become a ‘Reality Check’
The ETF sandbox may have gotten a little too wild.
The Securities and Exchange Commission announced last week that it will open its doors to public comments on “novel” ETF strategies in order to protect investors and foster innovation. The move is the latest by an agency that has been skeptical of highly leveraged strategies in recent months. The issue may be that the rules that worked well for some ETFs — namely Rule 6c-11, which lets ETFs operate under the ’40 Act without having to apply for exemptive relief — may not work for increasingly complex products.
“When products get too complicated, bad things happen,” said Adam Gana, a securities lawyer for Gana Weinstein. “Rule 6c-11 may be too flexible. It focuses heavily on ETF mechanics, but it does not really answer the harder question, which is: ‘Should every strategy that can technically fit inside the ETF structure be allowed to use that structure?’”
With global ETF assets recently surpassing $23 trillion, the agency’s action may be a little, well … too little, too late. Either way, it’s still a step in the right direction, said Amrita Nandakumar, president of Vident Asset Management. “The SEC’s review of so-called novel ETFs is a long-overdue reality check for an industry that may have pushed the boundaries of Rule 6c-11 too far,” she said.
Specifically, issuers may have taken advantage of the agency’s 75-day rule, under which a fund’s preliminary filing may contain generic language that the issuer can revise later. After the clock runs out, some providers modify the strategy significantly in the final prospectus, meaning the SEC can’t effectively regulate it until after it has begun trading. “Not only is it worth questioning whether 75 days is enough,” she said, “but also, are we entirely sure that some of these novel ETFs really do fall under [Rule] 6c-11, or should they be categorized differently?”
Nandakumar said some ETF categories that may now attract more agency scrutiny include:
- Private credit funds, which are subject to liquidity risk and price swings.
- Complex crypto strategies, which are costly and subject to closures.
- Single-stock derivatives, which don’t hold actual shares of a company, but instead use financial instruments like swaps to track an underlying stock and are subject to compounding decay.
Product Drift: So, what now? Given the SEC’s current Trump-appointed makeup, Gana doesn’t think there will be any kind of broad ETF rollback. What’s more likely to happen, he said, is that the agency will create a sharper distinction between traditional and “novel” funds, with more liquidity requirements, naming standards and possibly limits on certain products being marketed to retail investors — but that would be an “at best” outcome.
“ETFs began as efficient, transparent vehicles for diversified exposure,” he said. “Now we are seeing products tied to crypto assets, leverage, single-stock leveraged strategies, etc. Some of those may be appropriate. But they are not all the same from an investor-protection standpoint.”
The post SEC’s Interest in Novel ETFs Could Become a ‘Reality Check’ appeared first on The Daily Upside.
18:53
Cyclicality Beyond Crypto, Fading the Masses and More
Welcome Avatar! So far so good. We’re seeing another bear market rally and suspect that there will be at least one more leg down. Overall, we’re more focused on the social interest, on chain movements and time based capitulation vs. the monthly moves. Many people got sucked into BTC when it went from 70s to 90s again then 60s to 80 again (so on and so f…
19:23
100 A-Listers’ Summer Reads; A24-Google DeepMind Fears; NBCU Breakup
Happy birthday, USA! Here’s to making the most of this historic weekend, whether you’re glued to the World Cup drama or chilling poolside with a great book — and in case you’re still looking for that perfect summer read, we’ve got 100 for you to choose from, recommended by some of the biggest names in showbiz, media and beyond. Sony Pictures CEO Ravi Ahuja to Gov. Gavin Newsom, Jason Blum to Molly Shannon, Jay Ellis to Sherry Lansing, Amelia Dimoldenberg to Franklin Leonard all reveal what’s on their bedside table — and why.
Plus, we’ve also collected our top feature stories of the year so far — from Sean McNulty’s and Claire Atkinson’s excellent analyses of Paramount’s move to acquire Warner Bros. Discovery (and what might still get in the way) to Matthew Frank’s searing take on Hollywood nepo babies and Degen Pener’s memorable report on why cigarettes are making a comeback. Catch up on our 2026 hits!
And finally, as the industry grew increasingly quiet ahead of the holiday, Matthew made some noise with his scoop about an Academy shakeup — the exit of org veteran Teni Melidonian in a restructuring from CEO Bill Kramer. As AMPAS readies itself for the 100th Oscars and a transition from ABC to YouTube, expect more ripples ahead.
Now, ICYMI, the rest of our best of the week:
Save Our Sets: A Microdrama Lifeline Arrives — With Jobs Elaine Low toured the newly built sets at Sunset Las Palmas, where an L.A.-based microdrama studio is betting that vertical series — werewolves, steamy bedrooms and all — can buoy production.
Anne Rice, Zombies — and Netflix? AMC’s Plan to Outlast the Streaming Giants AMC’s chief content officer, Dan McDermott, tells Lesley Goldberg about the platform’s premium-for-adults strategy, Walking Dead plans and a new A24 sports show.
Vive L’NBCU! Hollywood works best when entertainment is the main event, says Richard — not a side hustle for a corporation. So, yes: He’s optimistic about Comcast spinning off NBCUniversal.
A24, Google DeepMind and the Dangerous Business of Selling Cool Much of Hollywood doesn’t just have questions about a new pact between the AI giant and the indie film stalwart — it wants it to fail. Erik Barmack analyzes the implications for both sides, but particularly A24 and its brand.
Christopher Nolan & the Short List of Directors Whose Names Open Movies The Odyssey sold out before a single review. Thank its director clout for that. Matthew reveals the three ingredients that turn a filmmaker into a box office draw today.
Hollywood’s Hopecore Summer: Doom on Pause As America turns 250 under a dark cloud, Katey Rich reflects on how culture — from the World Cup and the Knicks to Toy Story 5 and Project Hail Mary — is having a deeply optimistic moment.
Plus, Katey and Christopher Rosen check in on the coming Oscar race, with contenders like Project Hail Mary and Fjord already on voters’ minds:
Taylor Swift: First Comes Love, Then Marriage, Then Oscar in a Disney Carriage? In her quest for an EGOT, Taylor Swift is gunning for an Academy Award via her Toy Story 5 track. Rob LeDonne assesses the strategy.
Sean dives into the nitty-gritty of what NBCU’s split from Comcast will mean for both companies — and whether Netflix will follow its failed WBD merger with an NBCU pursuit:
Wednesday → Hollywood Stocks Q2 Winners & Losers
Tuesday → NBCU-NETFLIX & the Wild Card M&A in the Mix
Sunday → Supergirl, Jackass Fall Flat As Toy Story 5 Rolls On
Elaine, Natalie Jarvey and Sean dig into The Ankler’s inaugural summer reading list and unpack why book-based IP is hotter than ever:
Janice Min and Natalie break down who broke through in a week at Cannes Lions where subtlety was not the name of the game:
Can you explain 250 years of our nation’s history through movies? Richard and Chris debate the films that best capture the American experience, from 12 Years a Slave to Armageddon:
WSJ reporter Ben Fritz joins Sean to break down why Supergirl opened to just $38 million domestically, and what the result means for James Gunn and Peter Safran’s DC strategy:
New from Natalie Jarvey’s creator economy newsletter:
Summer House Scandal — and the Summer Reality TV Took Over
Andy Lewis’ latest IP picks:
24:04
Freemium: Can We Build Fair AI Models in 2026?
Hello!
Welcome to today’s edition of Business Analytics Review!
Here’s an uncomfortable truth we don’t say enough in this industry: a model can be 94% accurate and still be deeply unfair. Those two things — accuracy and fairness — are not the same question, and 2026 is the year most companies are being forced to reckon with that difference.
Think about it like a job interview panel that only ever hires from the same three universities. The panel might be “accurate” in the sense that it consistently picks capable people. But it’s also quietly, systematically shutting doors on everyone else. AI models trained on historical data do exactly this — except they do it at the scale of millions of decisions per second, in hiring, lending, healthcare, and insurance.
The good news? Fairness is no longer treated as a “nice to have” ethics footnote. It’s becoming an engineering discipline, a legal requirement, and — increasingly — a competitive advantage. Today, we’re digging into how companies are actually fighting the bias battle: through fairness-aware training and regular, structured audits.
For years, “we’ll fix bias later” was an acceptable answer. It isn’t anymore. Regulation has caught up fast, and it’s reshaping how seriously companies take this problem. The EU AI Act’s high-risk provisions move into full enforcement this year, with penalties reaching into the tens of millions of euros for non-compliance. New York City’s Local Law 144 has effectively become the template other US states are copying for mandatory bias audits on automated hiring tools. And regulators from South Korea to the EEOC are extending disparate-impact scrutiny straight into AI decision systems.
The upshot: fairness has moved from the ethics team’s slide deck to the procurement checklist. As one industry analysis put it plainly — the companies winning AI contracts in 2026 aren’t necessarily the ones with the smartest models. They’re the ones that can prove their models are fair.
Here’s where it gets genuinely interesting from a technical standpoint. Bias mitigation isn’t a single fix — it’s a layered strategy applied at three different points in a model’s life:
Before training (pre-processing) - This is about cleaning the fuel before it hits the engine. Teams audit datasets for imbalance, reweigh underrepresented groups, and increasingly use synthetic data to fill gaps where real-world data is skewed. If a facial recognition system only ever saw lighter skin tones during training, no amount of clever modeling afterward will fully fix that — the data problem has to be solved at the source.
During training (in-processing) - This is where fairness gets baked directly into the learning process. Techniques like adversarial debiasing pit two networks against each other — one trying to predict the outcome, the other trying to guess a protected attribute (like gender or race) from those predictions. If the second network succeeds too easily, the first one is penalized. Constraint-based optimization works similarly, treating fairness like a budget the model isn’t allowed to overspend.
After training (post-processing) - Even a well-trained model can produce skewed outcomes once deployed. Post-processing adjusts final predictions — recalibrating thresholds differently across groups — to bring outcomes back in line with a chosen fairness metric.
The catch, and it’s a big one: there is no universal definition of “fair.” Demographic parity, equalized odds, and counterfactual fairness are all mathematically legitimate — and often mutually incompatible. A model can’t satisfy all of them at once. This is why fairness has to be defined at the design stage, for the specific context the AI operates in, not bolted on as an afterthought.
Fairness-aware training only works if someone keeps checking the work. That’s the job of the bias audit — and increasingly, it’s not optional. One striking data point from recent industry research: 79% of organizations say fairness is a priority, yet only 24% actually run regular fairness audits. That gap is exactly where the risk lives.
A mature audit process looks less like a one-time compliance box-tick and more like a recurring health check: testing model performance across demographic groups, hunting for proxy variables that sneak protected characteristics in through the back door (a zip code standing in for race, for example), and repeating the process at every meaningful model update — not just at launch.
A real-world case that illustrates this well: a city government’s automated decision-making tool for a public services program was found, on audit, to be replicating bias baked into decades-old policy data it had been trained on. Rather than quietly patch it, the city paused the program entirely, retrained the model with explicit fairness constraints, and brought in an external advisory board to review it going forward.
29:31
Why AI is burning women out
Women spend twice as many hours per week as men on childcare and household work combined, and it starts before the alarm stops buzzing. A typical morning for me looks like waking the kids, discovering my son’s Spirit Week shirt still in the hamper, starting laundry, making breakfast I then don’t eat, realizing we’re out of cereal and adding it to the grocery list, and reminding myself to order a birthday gift, all while watching the clock for the 7:15 a.m. bus. Some version of this plays out for women everywhere, every morning, before they get to their desks.
Then the second shift starts: providing emotional support to a colleague before a big call, mentoring new teammates, leading an employee resource group meeting. Most of this work isn’t measured or rewarded, but it takes energy all the same.
On top of all of this, organizations are asking us to adapt to an entirely new way of working with AI. It’s no wonder conditions like “brain fry” and “thinkslop” are becoming an everyday experience. For many, the pressure of learning to use AI feels like adding another browser tab that never closes, but it’s impacting women more. In data from our 2026 report, Workforce State of Mind, we found over the past year that 73% of women say mental or cognitive strain has hurt their productivity, compared with 67% of men. Women are also more likely to report that strain is affecting their sleep quality (83% vs. 70% of men), their ability to focus (80% vs. 67%), and their engagement at work (69% vs. 59%). AI didn’t create this gap, but it’s making it wider.
Cognitive capacity is like a bank account, and AI adoption is a new monthly charge that’s the same amount for everyone. The mental overhead of prompting, fact-checking output, and applying results is taxing, especially wedged between meetings. Research shows it can take over 20 minutes to recover our focus when we go between disparate tasks.
But women, especially mothers, aren’t starting from the same balance. The domestic load doesn’t get redistributed just because both partners work. Women remain the default family managers, the go-to emergency contacts, appointment schedulers, and people tracking everyone else’s needs. At work, they’re still expected to be the emotional connective tissue of their teams, as the conflict diffuser, the morale keeper, and the one who notices when someone’s struggling. This is all layered onto an already demanding reality from just living in a woman’s body. Women spend 25% more time in poor health than men, driven by diagnostic delays and treatments that weren’t designed for female physiology. Hormonal shifts across the monthly cycle, through pregnancy and postpartum, and also into perimenopause can affect sleep, focus, and cognitive capacity in ways that are rarely addressed at work.
So when the same AI charge hits, it doesn’t land the same way. Men are often drawing from a surplus. Women are getting charged from an account that’s already overdrawn.
AI isn’t creating new disadvantages for women; it’s amplifying the ones that were already built into the system. Think of how competence is evaluated differently by design. Women face what researchers call the “prove it again” dynamic, where men are hired and promoted on potential, and women are judged on demonstrated proof of achievement. AI adoption gives organizations a new arena to apply that same standard, where women are starting from a higher baseline of expectations.
When a woman uses AI to produce strong work, the question becomes, “Did she really do that?” When a man does, it’s evidence of smart, efficient leadership. Women already lose credit for their contributions more readily. Among workers who have used AI on the job, men are 27% more likely than women to have been praised for doing so.
Women are also overrepresented in roles most exposed to early automation: the administrative, coordinative, and support functions that organizations are targeting first. This trend reflects decades of undervaluing work that women were tracked into, with women facing a higher threat of displacement at the same time they’re being asked to prove their proficiency with the tools doing the displacing.
And being watched while you learn carries its own weight. Women already face higher rates of imposter syndrome and workplace self-doubt. The pressure to prove they can keep up hits harder when they’re already questioning their place at the table. Adding AI fluency to the list has become its own kind of exhausting, stemming from a system that has always required women to perform competently for skeptical audiences. One-size-fits-all well-being programs weren’t designed for these unique stressors.
Organizations need to do more to support women. They can start by naming the gap directly and holding listening sessions where employees can be honest about what AI adoption is costing them, and where leadership shows up to hear it.
