CNBC’s latest SurveyMonkey “Women at Work” poll landed a clear and important signal: the generative-AI era is opening with a gendered split — men are more likely to call AI a “valuable assistant,” while women are noticeably more skeptical and far less likely to be frequent users at work. The difference is not marginal. The survey of 6,330 people (fielded Feb. 10–16, 2026) found that 69% of men say AI is a valuable collaborator versus 61% of women; roughly half of women said “using AI at work feels like cheating,” and far more women report
never using AI on the job. Those numbers matter because early adoption and familiarity with tools shape who wins promotions, who shapes workflows, and who gets invited into the rooms where decisions are being automated. What looks like a difference in attitude today can harden into an economic disadvantage for women tomorrow.
Background / Overview
Since ChatGPT’s debut in late 2022, generative AI moved from niche research demos into mainstream business tooling. Organizations now embed models into productivity suites, customer-service channels, code-assistant pipelines, and bespoke internal apps. The promise is large: faster writing and analysis, automated triage, creative ideation, and scale in knowledge work. The risk is equally large: bias baked into training data, inaccurate outputs used without oversight, and automation that changes job shapes faster than reskilling programs can keep pace.
Against this high-stakes backdrop, the SurveyMonkey/CNBC findings are not an isolated oddity. Multiple datasets and academic analyses from 2024–2026 show a persistent pattern: women use generative-AI tools less often than men, and women report more concerns about job displacement, bias, and the ethical implications of model outputs. The result is what experts now call an “AI gender gap” — a compound problem where representation, trust, access to training, and workplace power dynamics interact to produce widening career differences.
What the data actually says
Snapshot from the CNBC / SurveyMonkey poll
- Sample: 6,330 respondents; fieldwork Feb. 10–16, 2026.
- Core attitudinal gap: 69% of men describe AI as a “valuable assistant and collaborator”; 61% of women agree.
- Cheating perception: Roughly half of women say “using AI at work feels like cheating”; a notably smaller share of men share that view.
- Usage gap: 64% of women report never using AI at work, versus 55% of men. Power-user gap: 14% of men say they use AI “multiple times a day,” compared with 9% of women.
- Training and FOMO: More men report they want more training (59% of men vs. 35% of women say they fear missing out), while a larger share of women strongly disagree with the framing that not embracing AI equals missing out.
These figures sketch a two-speed adoption landscape: a cohort of early users (disproportionately male) who integrate AI into daily workflows, and a larger group (disproportionately female) that approaches AI cautiously or avoids it altogether.
Corroborating studies and broader evidence
Independent research and industry surveys echo the pattern. Large employer surveys, LinkedIn analytics and academic preprints from 2024–2026 repeatedly show lower adoption rates for generative tools among women and younger female cohorts expressing more caution toward AI’s risks. Workforce reports also highlight representation gaps in AI development teams and leadership roles — a context that colors how women perceive tools designed and promoted mostly by teams that don’t reflect them.
Why this gap exists: five interacting causes
No single cause explains the gender gap in AI adoption — it emerges from several linked dynamics.
1) Differential exposure and representation
Women remain underrepresented in technical teams that design, pilot, and evangelize AI at companies. When the people building and recommending tools don’t reflect the full workforce, rollout plans, training materials, and use-cases can miss the concerns and workflow realities of large employee groups.
2) Trust, risk perceptions and prior harms
Women, on average, report higher sensitivity to issues of fairness, privacy, and safety. Because AI systems have produced demonstrable harms (biased hiring filters, facial-recognition errors, skewed health recommendations), women’s higher skepticism often reflects realistic risk assessments rather than technology phobia.
3) Job-displacement risk concentrated in female‑dominated roles
Automation tends to threaten tasks, not genders. Yet many routine knowledge-work roles where automation can be most disruptive are disproportionately held by women (administrative roles, certain customer-service functions, middle-office workflows). The prospect of “cheating” or being replaced by a tool therefore reads as a tangible career risk.
4) Socialization and confidence gaps around self‑promotion
Multiple studies show women underreport skills, undervalue their expertise, and adopt new technologies more cautiously until they feel competent. AI’s reputation as a force multiplier exaggerates the cost of not being an early adopter, and early adopters often gain outsized advantages.
5) Access to training and sponsorship
The SurveyMonkey survey underlines an important nuance: men are
both more likely to use AI and more likely to say they need training. That suggests two simultaneous truths — men are more apt to take hands-on risks and to ask for formal learning, while women may be more likely to be excluded from early pilot programs or informal mentorship channels that accelerate practical competence.
The career consequences — why this matters beyond opinion polls
The danger is structural. Early adoption creates accumulated advantages: better productivity metrics, more visible deliverables, faster promotion tracks, and ownership of the AI‑driven workflows that shape downstream hiring and reward decisions. When a cohort of employees — by gender, race, or job level — remains less engaged with enabling technologies, the
“first mover” advantage compounds into long-term inequality.
Sheryl Sandberg framed the risk bluntly in a December interview: AI will be most challenging for people who don’t know how to use the tools, and early training access can determine who stays competitive. If men disproportionately get pilot access, mentoring and promotional lift from AI-enabled work, the existing “broken rung” at entry-level promotion can widen into a wider managerial and leadership imbalance.
At scale, firms risk creating a two‑tier workforce: AI‑fluent employees who accelerate up the ladder, and less-exposed employees who fall behind — with predictable impacts on pay, leadership diversity, and organizational resilience.
Corporate responses so far — promise and caution
Large companies and banks have publicly accelerated AI deployments. JPMorgan Chase’s recent investor‑day messaging made that clear: senior executives signalled major internal adoption, heavy tech budgets, and plans to redeploy staff into different roles. Banking leaders described internal LLMs and generative toolkits powering research briefs, client pitches, and automation of repetitive tasks.
That scale is instructive for two reasons. First, institutional adoption proves these tools aren’t hypothetical — they will materially change work. Second, the implementation choices banks and big tech make (who gets access, which roles are augmented or reduced, how redeployment is funded) set a playbook that other sectors tend to follow.
But the public rollout also reveals tensions: companies tout productivity gains while acknowledging displacement risks. Where retraining is an afterthought or where rollout metrics prioritize speed over inclusive adoption, the gendered adoption gap risks being baked into corporate practice.
Strengths in the current narrative — what’s hopeful
- Practical caution from skeptical users can prevent sloppy automation. Women’s higher scrutiny on bias, privacy, and fairness can guard organizations against deploying brittle or discriminatory systems.
- Growing public attention to the gap is generating concrete responses: diversity‑targeted training, corporate dashboards that track AI usage by demographic cohort, and nonprofit initiatives aimed at closing digital skills gaps.
- High‑profile warnings and leadership commentary (from CEOs and former industry executives) are spurring boards and HR leaders to take reskilling seriously rather than assuming technical change is benign.
These dynamics suggest the gender gap can be narrowed before it calcifies — if companies design adoption strategies around inclusion rather than speed.
Risks and unresolved questions
- Asymmetric rollout: If pilot programs privilege certain teams or geographies, the adoption gap will widen and hard-to-repair career inequalities will grow.
- Measurement blind spots: Few companies currently publish adoption metrics disaggregated by gender, race, or level. Without measurement, inequities remain invisible.
- Retraining fidelity: “Retraining” is often a catchphrase. Real-world retraining requires time, career pathways, mentorship, and credible promotion signals; superficial short courses won’t suffice.
- Trust deficit: If workers view AI as dishonest (the “feels like cheating” reaction), enforcement policies or permissive deployment can trigger pushback, compliance issues, and reputational cost.
- Bias in models and tooling: Tools trained on biased corpora will reproduce past inequities. Deploying such tools without robust model evaluation and guardrails will amplify harms.
Where numbers vary across reports — for example, internal counts of how many employees use an LLM at large enterprises — public reporting is incomplete and sometimes inconsistent. That inconsistency itself is a risk: opaque data makes it harder for regulators, unions, and civil-society groups to advocate for equitable deployment.
Practical playbook — how companies should close the AI gender gap now
The fix is not a single program; it’s a multi-year strategy across policy, learning, and culture. Below are concrete steps organizations should adopt immediately.
1. Measure adoption and outcomes
- Track AI use frequency, by role, level, gender, and geography.
- Report anonymized, aggregate dashboards to leadership quarterly.
- Tie adoption metrics to promotion and performance analytics to spot early skew.
2. Democratize access to pilots and early tools
- Open AI pilots across representative cross-sections of teams, not just technical centres of excellence.
- Use randomization in pilot selection where possible to avoid sponsorship bias.
- Remove hidden barriers (admin rights, workspace enrollment) that favor certain groups.
3. Build role‑specific training paths
- Develop task‑based learning — show how a tool helps with specific job tasks rather than offering generic demos.
- Offer hands-on labs, mentoring with AI power users, and micro‑credentials that inform promotion decisions.
- Fund time for learning (protected hours), not just one-off lunch-and-learns.
4. Make AI literacy part of performance and promotion criteria
- Recognize AI‑enabled productivity in evaluations so early adopters don’t get an unearned advantage while late learners are penalized.
- Ensure promotions account for a candidate’s willingness to reskill and coach others, not just raw output.
5. Invest in trustworthy AI and transparency
- Require model cards and explainability tests for internal tools.
- Establish human-in‑the‑loop approvals for decisions that materially affect people.
- Run fairness audits and make remediation a condition for deployment.
6. Address trust and ethics head-on
- Communicate transparently about what AI should and shouldn’t do.
- Create channels for employees to flag errors, harms, or biased outputs without fear of penalty.
- Include skepticism as a useful lens — surface concerns and act on them.
7. Sponsor women into AI roles and leadership
- Actively recruit women into AI product, governance, and data science roles.
- Provide visible sponsorship (not just mentorship) for women leading AI initiatives.
- Create internal sabbaticals or rotation programs to broaden experience in AI product teams.
For policy makers and industry groups
Public policy can amplify corporate action. Governments and standards bodies should:
- Support funded retraining programs targeted at sectors where women are overrepresented.
- Require larger employers to report AI adoption and retraining outcomes disaggregated by demographics.
- Fund independent audits of expensive, high-impact internal AI systems (hiring, lending, legal decisions).
- Encourage standards for model transparency and portability so organizations can swap to less-biased vendors.
These actions create a level playing field and make it easier for companies to commit to inclusive practices without competitive disadvantage.
What employees and individual contributors can do today
No single worker is responsible for systemic adoption patterns, but individuals can still take practical steps:
- Experiment safely: start with sandboxed prompts and share learnings with peers.
- Document time saved and outcomes produced using tools to build evidence for reskilling time.
- Seek mentors who use AI tools in domain-specific ways, not generic demos.
- Advocate for transparent pilots and equitable access within teams.
Collective employee voice matters: labor groups, affinity networks, and employee resource groups can accelerate company action by negotiating training access and accountability mechanisms.
The longer view: closing the loop between ethics and adoption
The gender gap in AI adoption is both a symptom and a signal. It reveals a deeper mismatch between how AI is built and how work actually gets done across diverse teams. The good news is that skepticism is not a flaw to be corrected by pressure; it’s information that can make deployments safer and more durable.
To avoid a future where AI cements existing disparities, leaders must translate skepticism into design criteria: fairness, transparency, and inclusive rollout. When organizations treat trust and access as product metrics as important as latency and accuracy, they create systems that benefit more people and are less likely to cause harm.
Final assessment: risk, responsibility and opportunity
The SurveyMonkey/CNBC results are a wake-up call. If nothing changes, differential adoption will amplify career gaps; if companies heed the evidence and act strategically, they can use this moment to build more inclusive workplaces and fairer technology.
- Risk: Left unchecked, early adoption advantages will concentrate in a subset of employees and managers, worsening promotion and pay gaps.
- Responsibility: Employers must measure and mitigate biases in access and outcomes; regulators and civil society must press for transparency and fairness.
- Opportunity: If training, pilot access, governance, and product design become explicitly inclusive, organizations stand to gain broader buy‑in, safer deployments, and stronger long-term productivity gains.
AI is a tool that will change work. Whether it narrows or widens existing gender gaps depends on decisions we make now: who gets access, how we teach, what we measure, and whose voices we elevate in the design and deployment of the systems that will run our businesses tomorrow. Confronting the gender gap in AI is not merely a diversity objective — it’s a business imperative and a test of whether the AI era will deliver equitable productivity for everyone.
Source: CNBC
AI's got a gender gap: Women are more skeptical