Is AI Recruitment Tool Hindering Women's Careers? The Era of 'Beautifying' Resumes

To present oneself in the best light during a job search, refine the text of the resume. Remove unnecessary career history and emphasize achievements relevant to the position being applied for.

Up to this point, these are quite standard job-hunting strategies.

However, in the UK, there are middle-aged job seekers taking even more drastic measures.

They are "removing information that could hint at their age."

Stacey Duguid, who has held senior positions in the fashion and publishing industries in the UK, began job hunting again in her 50s, considering returning to organizational work.

However, despite never having struggled significantly in her career search before, she found herself sending out a large number of resumes over 16 months, only to receive automated replies and not even progressing to interviews.

This led her to start what she describes as "botoxing her CV."

She reduced information that might suggest her age from her resume and shortened her lengthy career history. She trimmed old career details so that it wouldn't be immediately apparent how many decades she had worked.

The irony is that she had to hide what should have been a strength—her "long experience."

When she shared her experience on social media, she received numerous responses from women saying, "I'm in the same situation."

Is the issue merely the difficulty of changing jobs, or is it that AI-based recruitment systems are unknowingly reproducing existing biases linked to "age" and "gender"?

Applied to over 400 companies, with little response

To investigate this issue, the BBC spoke to over 60 women aged between 40 and 65.

Many were women who had worked for years in managerial or professional roles.

A notable feature is that many of them were not previously in a position where they had to look for work.

Koeyli Jaluka, 49, is one such person.

She has held senior positions in manufacturing, corporate divisions, and healthcare, and for over a decade, she was often approached by companies without having to apply herself.

However, after losing her job, the situation changed dramatically.

She applied for over 400 jobs, yet only a few companies responded concretely.

She was sometimes told she was "too senior."

She felt this phrase was another way for companies to say, "You're too old," without directly stating it.

Some women are advised by career coaches to "cut the first 10 years or so of their career from their resume."

In other words,

it's not about increasing experience, but about reducing it to improve job prospects.

Such a reversal is beginning to occur.

Anna Cowie, 52, has also worked in the advertising industry for about 30 years. However, in recent years, she has been unable to secure a job.

Despite having experience with major brands and participating in large redevelopment projects in London, she has applied so many times she has lost count.

What she finds more promising is not the method of mass online applications.

Instead, it's volunteering at local events and connecting with real people.

This allows her to be seen as a person, not just a string of text on a resume.

This sentiment symbolizes the issues present in the current hiring market.

AI can infer "age" even without being provided with it

The important point here is that many AI recruitment services do not necessarily set "age" or "gender" as direct evaluation criteria.

Companies likely do not introduce AI with the intent to discriminate.

Yet, biases can still occur.

For example, if a resume includes,

"25 years of managerial experience,"

"Joined in the 1990s,"

"7-year career gap,"

"Worked in the same industry for over 20 years,"

even without the age itself, it's not difficult to infer the candidate's approximate generation.

Moreover, if AI is learning "good candidate" patterns from data of previously hired individuals or vast information on the internet, it might also incorporate societal biases.

The issue then becomes,

"Is AI discriminating," or is it "learning human societal discrimination"?

This is the question.

It's likely impossible to completely separate the two.

Stanford research reveals AI's assumption that "women are young"

Research supporting this concern has emerged.

Researchers from Stanford University and others examined how online images, videos, and large language models associate occupations with age and gender.

The research team used 54 types of occupations and typical male or female names to generate over 34,500 virtual resumes with ChatGPT.

It was found that resumes created with female names tended to depict individuals as "younger and less experienced" compared to those with male names.

Furthermore, in experiments evaluating the generated resumes, older men were reported to be rated more highly, even when starting from the same conditions.

It's important to note that this research does not prove that "all recruitment AI used by companies is rejecting middle-aged women."

It examined the social stereotypes embedded in general generative AI.

However, if similar models or concepts are used for resume creation, candidate searches, scoring, or recommendations, the risk of such biases entering the hiring process cannot be ignored.

AI does not create biases out of nowhere.

It learns extensively from values present in the internet, texts, images, and past human judgments.

It absorbs implicit images like "managers should be male" or "younger women are preferable," which society has held for years, and could reproduce them rapidly and on a large scale.

AI could become a device that reproduces these biases quickly and on a large scale.

The risk of "career gaps" being processed as a lack of ability

Career gaps, which particularly affect women, are a concern.

Many women temporarily leave work due to childcare, caregiving, or family reasons.

A human recruiter might understand,

"This period was for childcare,"

"There is 10 years of managerial experience before that,"

"There are sufficient skills available upon return,"

and read the background of the resume.

However, in machine screening, career gaps might be processed as mere "deficiencies" or "red flags."

Life circumstances that are explainable to humans become data points that lower scores for machines.

These issues are not limited to women.

They relate to anyone with a "non-linear career," such as those returning from illness, caring for family, failing in entrepreneurship, working freelance for long periods, or living abroad.

As AI recruitment expands, only those with "clean career histories" might be easily understood by algorithms, potentially disadvantaging complex life experiences.

It's also dangerous to conclude that "AI alone is the cause"

However, it's not possible to blame the current tough job market entirely on AI.

There are multiple factors simultaneously at play, such as reduced recruitment budgets, economic uncertainties, increased applicants per job, expanded competition due to remote applications, and more specific company requirements.

For the women interviewed by the BBC, it's difficult to individually prove that they were "rejected by AI."

In the first place, companies often do not explain to applicants,

which AI service was used,

at what point it was automated,

what score led to rejection,

or who made the final decision.

From the job seeker's perspective,

it's only visible that,

"I applied,"

"I received a rejection email immediately,"

or,

"There was no contact."

This lack of transparency breeds distrust.

It's unclear whether AI selection is the issue, whether a recruiter read and rejected the resume, or whether the job itself was effectively on hold.

The situation of "not knowing why one was rejected" might be one of the biggest issues with AI recruitment.

Jobs lost to AI and automation, while 12,100 digital roles remain unfilled

The issue isn't just about recruitment.

The City of London Corporation's Women Pivoting to Digital Taskforce points out that many women's jobs could be replaced by AI and automation by 2035.

Meanwhile, the UK faces a severe shortage of digital talent.

According to the task force, 12,100 digital roles in finance, professional services, and technology went unfilled in 2024 alone.

In other words,

"While some people may lose jobs to AI, companies are struggling with a talent shortage."

This contradiction is occurring.

Furthermore, a survey of over 1,000 women found that 89% of those interested in transitioning to the digital field would be willing to undergo retraining if offered opportunities by their companies.

However, only 32% were actually provided with opportunities for retraining or transitioning to digital roles by their current employers.

This means about 68% were not given such opportunities.

Companies may claim "a lack of digital talent" while not sufficiently investing in retraining existing staff.

This is why the reemployment issue for middle-aged women cannot be dismissed as merely a "personal job-hunting skill" problem.

Lawsuits over AI recruitment are also beginning

In the US, issues surrounding AI recruitment tools have reached the courts.

In a lawsuit involving HR software giant Workday, claims are being contested that its AI-powered recruitment tools unfairly excluded applicants based on race, age, disability, and other factors.

In June 2026, a US federal court ruled that some claims could proceed in the lawsuit.

However, this does not mean the court recognized that "Workday's AI engaged in discrimination."

The trial is ongoing, and the plaintiffs' claims have not been ultimately accepted.

Workday denies the allegations.

The company explains that AI does not make hiring decisions but serves as a support tool that shows recruiters how applicants' qualifications and experiences match job requirements.

They also state that age, gender, and other protected attributes are not used in decision-making, and they conduct ongoing fairness testing of their AI.

This rebuttal is also important.

Using AI does not necessarily increase discrimination.

Human interviewers, of course, also have biases.

It's long been pointed out that judgments can unconsciously change based on name, appearance, manner of speaking, alma mater, age, gender, and more.

Properly designed, audited AI combined with human checks could potentially reduce such human biases.

The issue is,

it's not "AI or human," but "how it's used."

That's where the focus should be.

On social media, there's empathy with "I'm the same," while recruiters also express dissatisfaction with AI

The reactions on social media symbolize this issue.

When Duguid shared her job search experience, voices from women of similar age expressing similar experiences gathered.

In subsequent LinkedIn posts, she argues that the issue isn't just AI but that age bias exists throughout the recruitment process, from reading resumes to narrowing down candidates, inviting them for interviews, and whether companies respond.

Another LinkedIn post points out that regarding unemployment and reemployment of women in their 50s, companies might be assuming "older people might demand higher salaries," "they might be weak in digital skills," or "they might lack flexibility."

On the other hand, looking at recruitment-related communities on Reddit, questions about AI recruitment tools arise not only from job seekers but also from recruiters.

A user who has been in recruitment for years questions whether adding AI interviews or automated screenings truly saves time, as recruiters ultimately need to see candidates.

In another discussion, while AI is seen as useful for tasks like scheduling, record-keeping, and template emails, concerns about explainability and bias remain for important decisions like candidate evaluation and ranking.

##HTML_TAG_294