Womafreesm is launching an international audit to examine how
markets, hiring practices, and AI systems undervalue
women aged 45-64.
If these systems are wrong, the error ceases to be a private stereotype. It becomes a mechanism that affects budgets, careers, visibility, trust, and access to opportunities.
The World Organization of Women Nations Womafreesm is conducting an international comparative audit to measure this pattern and create a practical framework for identifying it.
International Audit of Systematic Misclassification and Underrepresentation of Women Aged 45-64 in Market and AI Systems.
Marketing, recruitment, and AI systems.
Visibility, authority, relevance, and perceived value.
Germany, Poland, Spain, Ukraine, United Kingdom.
Women aged 45-54 and 55-64 as two separate analytical groups.
December 2026.
Women aged 45-64 aren’t just “underrepresented.” They may be misclassified by the systems themselves.
For too long, the situation of women aged 45-64 has been described in cultural terms: “they’re overlooked,” “they’re pushed aside,” “they’re undervalued,” “the market favors youth.” But today, that language is no longer enough.
Because decisions about visibility, status, and opportunities are increasingly made not just by people.
They are made by markets, platforms, HR systems, algorithms, generative AI tools, and data-trained models where old stereotypes are already embedded in the language, imagery, and evaluation criteria. Once such a stereotype enters the system, it ceases to be a personal opinion.
It begins to scale.
What was once one person’s prejudice becomes a repeatable logic of selection, segmentation, description, and recommendation.
Because a misjudgment leads to poor decisions
A woman aged 45-64 can be a financially capable consumer, a strong candidate, a leader, an expert, an entrepreneur, a mentor, and a source of rare expertise. But the system may see her differently.
In marketing, she may fall outside the image of a valuable audience: the one for whom products, budgets, and campaigns are created.
In hiring, her experience may start working against her. Instead of maturity and competence, the system sees risk: “too old,” “too expensive,” “will have a hard time fitting into the team.”
In AI descriptions, her authority may be diminished at the level of wording: the system describes her as softer, weaker, or less expert than a man with comparable experience.
In professional language, her maturity may lose its status. What is often interpreted as gravitas, experience, and reliability in a man can become an age-related disadvantage for a woman.
This is no longer a matter of image.
It is a matter of the quality of decisions:
When the system misjudges a woman, it is not only wrong in its assessment of her. It allocates attention, money, opportunities, and trust less effectively.
In 2025, the journal Nature published one of the most extensive empirical studies on age and gender bias in online media and language models.
The analysis covered nearly 1.4 million images and videos from major online platforms, as well as nine language models. The study found that women are systematically portrayed as younger than men across 3,495 professional and social categories. This gap is particularly noticeable in high-status and high-paying professions.
A separate experiment with resumes showed that when generating and evaluating professional profiles, ChatGPT attributed a younger age and less experience to women, while assessing older men as more qualified.
This study identified an important pattern: age and gender are distorted not only in culture but also in digital systems, which are increasingly involved in evaluating people. Womafreesm is taking the next step: we are measuring how this pattern manifests in a specific group of women aged 45-64, across specific decision-making domains, and in different countries.
Where does the stereotype end and the systemic
bias begin?
We are investigating whether there is systematic misclassification of women aged 45-64 in markets, hiring, and AI systems.
By misclassification, we mean a consistent pattern in which women aged 45-64 are attributed lower relevance, authority, or value without a proven reason.
In other words, the system rates them lower than men of comparable age and qualifications or younger women, even when actual data on competence, purchasing activity, or professional contribution does not support this.
Simply put:
The system may see a woman, but not consider her significant.
It may display her, but not attribute authority to her.
It may formally take her into account, but exclude her from the segment, budget, shortlist, or expert field.
When the system misjudges a woman, it is not only wrong in its assessment of her. It allocates attention, money, opportunities, and trust less effectively.
One of the mistakes made when discussing women aged 45-64 is that everything is reduced to a single question: “Are they being shown or not?” But visibility alone doesn’t prove anything. A group can be visible yet lack influence. It may appear in advertising but be excluded from the premium segment. It may be mentioned in hiring, but not make the shortlist. It may appear in an AI response, but be described as weaker, softer, and less expert.
Whether women aged 45-54 and 55-64 appear in descriptions of clients, candidates, experts, and audiences.
What words are used to describe their competence, experience, status, trustworthiness, and leadership.
Whether they are considered a significant audience, strong candidates, and valuable market participants.
These three layers should not be confused. Visibility does not necessarily imply authority, and authority does not always imply recognized market value. Therefore, Womafreesm analyzes them separately to identify where exactly the undervaluation occurs: in presence, language, status, or the distribution of opportunities.
Every day, brands decide:
If women aged 45-64 are absent from marketing materials or represented only narrowly, the market receives a distorted picture of their economic power. That is why Womafreesm analyzes marketing as the space where the image of the valuable consumer is formed.
Women aged 45-64 often stand at the intersection of experience, responsibility, and influence.
By this stage of life, many have built careers, managed families, navigated institutions, made financial decisions, and developed the ability to recognize patterns, risks, and long-term consequences.
They carry professional expertise, purchasing power, social responsibility, and a level of maturity that shapes how communities, workplaces, and households function.
But it is precisely at this age that the system increasingly begins to view women through a lens of decline:
This creates a gap between a woman’s real value and how she is described, portrayed, evaluated, and included in decisions. This is more than just a cruel cultural habit.
It may be an economic and algorithmic error: the market, hiring practices, and AI systems may undervalue a group that actually possesses experience, money, influence, and the ability to make decisions.
This is precisely why Womafreesm divides women aged 45-54 and 55-64 into two analytical groups. Their positions may differ in terms of visibility, authority, and perceived value. If they are lumped into a single category, an important part of the picture may be lost.
Existing research already reveals age- and gender-based biases.
However, three critical gaps remain in this area.
Women over 45 are often grouped into a single broad age
category. But for the market, hiring, and professional status, the
difference between 45-54 and 55-64 can be fundamental: these
perceived, described, and evaluated differently.
The professional context is analyzed more frequently. But it is
marketing that shapes the image of a valuable consumer every
day: who the market wants to see, to whom the product is
addressed,
and whose purchasing power it recognizes.
Public job postings reveal only the official language. Employers’
real expectations often emerge later: in long lists, shortlists,
comments on the ideal profile, and informal age markers.
we analyze the 45-54 and 55-64 age groups separately, combine marketing, hiring, and AI systems, and examine not general impressions but recurring mechanisms of undervaluation.
This study is structured as an international comparative audit with a mixed-methods design: we compare data from different countries, different sources, and different policy scenarios. A pattern is considered significant only if it recurs in at least three independent data sources.
A single quote may be a coincidence. A single expert comment may reflect personal experience. We are interested only in recurring patterns: how different systems repeatedly describe, evaluate, and distribute the value of women aged 45-64.
We look beyond whether women aged 45-64 are present in advertising, hiring, or AI responses. We examine the status attributed to them: strong, weak, relevant, secondary, expert, or questionable.
The study takes into account language, imagery, AI responses, professional markers, marketing materials, and real-world market practices. This allows us to see not just isolated signals, but a mechanism that can be repeated across different systems.
A structured experiment using comparable prompts for leading generative AI systems.
We are examining how these systems describe customers, candidates, and experts when given the same data but different age and gender parameters.
Goal: To determine whether AI diminishes the authority, relevance, and value of women aged 45-64 at the linguistic level.
Content analysis of advertisements, landing pages, audience descriptions, and visual materials in categories where women aged 45-64 are actual consumers.
We examine who the market portrays as a desirable, solvent, and significant audience.
Objective: to identify where the economic power of women aged 45-64 disappears from marketing communications.
Structured engagement with recruitment agencies and analysis of aggregated, anonymized data.
We examine how age and gender may influence employer expectations, long lists, shortlists, and comments on suitable profiles.
Goal: to identify the hidden logic of selection that is rarely visible in public job postings.
Research of job postings, corporate biographies, industry media, and public descriptions of expertise.
We examine which words are associated with experience, leadership, reliability, flexibility, relevance, and professional standing.
Goal: to understand when a woman’s maturity is perceived as authority and when it becomes an age-related risk.
Comparison of descriptions of women aged 45-54, women aged 55-64, men aged 45-64, and women aged 25-39 in similar professional and consumer contexts.
We examine whether the language of evaluation changes when experience, status, and role are comparable.
Goal: to distinguish real differences in the data from age- and gender-based bias.
30-40 interviews with marketers, HR directors, brand strategists, AI developers, and market researchers.
We validate the identified patterns through the professional experience of people who work with audiences, hiring, brands, and AI tools.
Goal: not to replace data with opinions, but to clarify the context and practical logic of the market.
The study has already moved from the conceptual phase to the operational phase.
Since November 2025, conceptual preparations have been underway: the central research question was formulated, the target age group (45-64) was specified, key areas of analysis were identified, and the study's connection to the Womafreesm Evidence Data program was established.
Since January 2026, the study has entered its active phase: the design of the first phase was finalized, countries were selected, analysis domains were defined, and the first research tools were prepared.
At this stage, the research is no longer confined to a theoretical framework. It is moving on to the collection of initial field and experimental data: through recruitment agencies, AI audits, and the preparation of a comparative analysis by country.
The study will yield an evidence base that can be used not only to describe the problem, but also to test systems across marketing, hiring, AI, media, partnerships, and institutional decisions.
The study should provide the market and society with a new language: how to identify, measure, and prove the underestimation of the relevance of women aged 45-64.
for foundations, universities, institutional partners, companies, and professional communities.
for visibility, authority, relevance, and perceived value.
for brands, employers, recruitment agencies, and AI developers.
with separate data on women aged 45-54 and 55-64 in the countries of the first phase.
for journalists, donors, researchers, and strategic partners who need a quick and accurate introduction to the topic.
This study provides precise language for a topic that is often described in overly vague or general terms.
Not “women over 45 are underrepresented.”
But “systems may underestimate their relevance.”
Not “the market favors the young.”
But “market segmentation may mistakenly exclude an economically active group.”
Not “AI is biased.”
But “AI may reproduce a specific age-gender pattern in hiring, marketing, and expertise.”
For journalists, this shifts the topic from an opinion piece to a verifiable social issue: one with data, language, comparisons, and a clear mechanism.
Foundations and institutional partners don’t need slogans; they need issues that can be substantiated, measured, and translated into a practical framework.
This study brings together several fields that are typically analyzed separately:
Its value lies not only in describing the problem. Its value lies in creating a framework that helps identify, compare, and then correct recurring age-gender patterns.
For partners, this serves as an entry point into a topic that combines social significance, methodology, and the potential for future practical application.
If women aged 45-64 are systematically classified as less relevant, the market loses more than just fairness.
For the market, this is not an abstract ethical issue. It is a matter of the quality of decisions: who a product is targeted at, who is hired, who is trusted, and where the system mistakenly fails to recognize value.
Because for far too long, women’s maturity has been described as a decline.
This study reframes the question itself.
Not “why does a woman become less valuable with age?”
But “why does the system begin to perceive her value less accurately?”
For women aged 45-64, this is not a text about defending their image. It is an attempt to prove that the problem may not lie in their age, experience, or relevance, but in the accuracy of the systems that evaluate them.
Womafreesm is launching this study at a time when the market, funds, media, employers, and AI developers already need a new tool: not just to talk about bias, but to show exactly where the system goes wrong, whom it underestimates, and how this can be verified.
The research is open to foundations, donors, universities, research centers, recruitment agencies, brands, AI teams, professional networks, and media outlets that are committed to shaping a new agenda on age, gender, the labor market, and AI.
Supporting the research means participating in the creation of the first practical layer of an audit that can demonstrate how systems underestimate the relevance of women aged 45-64 in marketing, hiring, and AI.
Can request a donor package and provide financial support for the research. A presentation, budget rationale, timeline, research modules, and participation options have been prepared for them.
Can provide aggregated and anonymized data on how women aged 45-54 and 55-64 navigate real-world hiring filters: employer expectations, long lists, shortlists, and comments on suitable profiles.
Can participate as professional market partners: provide examples of communications, segmentations, campaigns, or expert commentary on how women aged 45-64 are represented as consumers, clients, and audiences.
Can participate in a professional dialogue on how language models describe age, authority, expertise, and professional value in different scenarios.
Can help bring the research to HR directors, marketers, brand strategists, recruiters, market researchers, female entrepreneurs, and women aged 45-64 with strong professional experience.
Can request a press release, interview, or comment from Womafreesm for coverage on age, gender, AI, the labor market, marketing, and women’s economic visibility.
Can join as academic and research partners: participate in expert discussions, international dissemination of results, scientific publications, conferences, and the next wave of research.
A donor package with details is provided to foundations, major donors, and institutional partners upon request.
The world is entering a period when AI systems can no longer be evaluated solely on the basis of convenience, speed, and efficiency.
The next question is already being framed differently: whom do these systems make visible, whom do they marginalize, whom do they exclude, and what decisions do they distort. In the European Union, the AI Act is already establishing a risk-based framework for AI systems, including those used in employment, recruitment, and workforce management.
This means that AI in hiring and professional assessment is gradually shifting from the realm of experimentation to one of accountability, documentation, and oversight. (digital-strategy.ec.europa.eu)
In the UK, the ICO has already reviewed AI tools in recruitment and issued recommendations for developers and users of such systems regarding fairness, transparency, and the protection of candidates’ data. This shows that hiring is becoming one of the first areas where AI risks will require evidence, not just slogans. (ico.org.uk)
In its AI risk management framework, NIST specifically highlights the harm associated with bias, discrimination, and the reinforcement of historical and social distortions in generative AI systems. In other words, the question is no longer “does AI have bias?” The question is who can identify, measure, and mitigate it. (nvlpubs.nist.gov)
This is precisely why it is now important to examine not abstract “AI bias,” but specific groups, specific scenarios, and specific consequences.
Women aged 45-64 are at particular risk: they intersect across age, gender, the labor market, consumer power, expertise, and digital representation.
If this pattern is not measured now, it may be embedded in new evaluation systems as “normal logic”: in marketing segmentation, recruitment filters, AI descriptions, professional ratings, and future automated verification standards.
This audit is part of the Womafreesm Evidence Data program. This program is designed to translate women’s experiences from personal stories into evidence-based data, measurable indicators, audit frameworks, and institutional decisions.
Womafreesm Evidence Data addresses the central question: What do markets, employers, AI systems, medicine, the economy, and institutions lose when women’s realities are not taken into account in the design of solutions?
The audit of women aged 45-64 is the program’s first major international research module. It demonstrates how a specific group of women can become less visible, less authoritative, and less relevant in systems that allocate attention, trust, budgets, career opportunities, and expert status. This audit also establishes the first evidence base for Womafreesm’s flagship product:
The WLCE Index will measure how much the economy, healthcare, the labor market, and institutions lose when they design systems without accounting for the female life cycle: age, body, health, reproductive stages, caregiving burdens, career transitions, and women’s social roles.
The current audit takes the first step in this framework: it examines how women aged 45-64 are categorized in marketing, hiring, and AI systems, and where exactly the system begins to undervalue them.
Womafreesm is building an international infrastructure for women’s influence through research, data, legal pathways, partnerships, and institutional solutions.
This audit is part of the Womafreesm Evidence Data program and is linked to the organization’s future flagship product, the Women’s Life-Cycle Economy Index.
The research is led by the Womafreesm team, which works across several areas: research design, collaboration with recruitment agencies, AI auditing, data analysis, partnerships, and international communications.
Marina Sukhomlinova
Founder & CEO of The World Organization of Women Nations WomafreesmSeparate materials detailing the research, stages, modules, and participation formats have been prepared for donors, foundations, universities, media, and institutional partners.
Yes. The conceptual phase began in November 2025. The study entered its active phase in January 2026.
Currently, the countries for the first phase, the areas of analysis, and the age subgroups 45-54 and 55-64 have been identified; research tools have been prepared; prompts for the AI audit have been developed; data specialists have begun working with language models; and Womafreesm has started working with recruitment agencies to collect data.
The first version of the international comparative report, a public executive summary, and the first practical audit framework.
These materials will show how women aged 45-64 are categorized in marketing, hiring, and AI systems, and where their visibility, authority, relevance, and perceived value may be underestimated.
An expanded academic version and partner distribution are planned for the first quarter of 2027.
Because these groups may be in different positions regarding visibility, authority, and perceived value.
Women aged 45-54 and 55-64 may be described differently by the market, perceived differently by employers, and classified differently by AI systems. If we combine them into a single broad category, important differences may be lost.
Because a woman’s value today is not determined in just one place.
Marketing decides who to consider a desirable audience.
Hiring decides who to consider a strong candidate.
AI systems are increasingly involved in describing profiles, segments, resumes, experts, and professional value.
If the same pattern repeats across multiple spaces, it is no longer a single stereotype. It is systemic logic.
Because public job postings reveal only the official language.
The real logic of selection often emerges later: in employers’ expectations, long lists, shortlists, comments on suitable profiles, and hidden age markers.
Recruitment agencies help reveal this layer of the market without sharing candidates’ personal data.
No. The recruitment module is built on aggregated and anonymized data.
Womafreesm does not request personal resumes, candidates’ names, contact information, or private details. The research is not interested in the fate of an individual, but in recurring patterns within the selection system.
The study can be supported by foundations, donors, universities, research centers, recruitment agencies, brands, AI teams, professional networks, and media outlets.
Foundations and donors can request a donor package.
Recruitment agencies can participate in the collection of aggregated data.
Brands and marketing teams can participate as professional market partners.
Journalists can request a press release, interview, or comment from Womafreesm.
The financial estimate is not published on the study’s public page.
A separate donor package detailing the budget, phases, research modules, and support options has been prepared for foundations, major donors, and institutional partners. It can be requested directly from the Womafreesm team.