The Department for Work and Pensions announced recently that around 20,000 people had used its Work Hub CV builder since June, while 16,000 had used its AI-powered Work Assistant for help with job searches, applications and interviews.

Helping jobseekers present themselves more effectively is a reasonable objective. But the announcement risks missing the more important change taking place in recruitment.

AI is doing more than helping people write better CVs. It is steadily devaluing the CV as a signal of ability.

When every candidate can produce a polished document tailored precisely to a job description, the document tells an employer less about the person behind it. DWP may be helping applicants navigate today's recruitment process while employers are already placing more weight on other ways of proving capability.

When every CV is polished, polish stops being valuable

CVs have always been imperfect. They are self-reported summaries that reward confidence, fluency and familiarity with recruitment conventions as much as underlying capability.

Nevertheless, they once contained useful signals. Candidates made choices about what to include, how to describe their experience and why they appeared relevant to a particular role. A recruiter could learn something from those choices.

Generative AI weakens that signal. A job description can be fed into an AI tool alongside someone's employment history, producing a CV that mirrors the employer's language, reproduces the desired competencies and is optimised for automated screening.

This does not necessarily involve dishonesty. Many applicants use AI sensibly to overcome poor writing skills, identify transferable experience or avoid being rejected by crude keyword filters. But when everybody can achieve the same level of polish, polish ceases to distinguish anybody.

Michael Page's 2026 Talent Trends research, based on almost 60,000 responses, found that 67 per cent of UK jobseekers use AI during the application process. It also found that 66 per cent of UK organisations have adopted a skills-based approach to hiring. Qualifications and career history remain relevant, but employers are putting greater emphasis on evidence of capability.

The consequences are already visible elsewhere. A US survey by Robert Half found that 65 per cent of hiring managers thought AI-enhanced applications made candidate skills harder to verify. Employers were responding with closer scrutiny, additional interviews and targeted skills assessments.

Having spent much of my career in recruitment, I understand why. A CV can tell me what somebody claims to have done. It cannot, by itself, show how well they did it, whether they can repeat it or how they would perform in a different environment. AI has widened that gap between presentation and proof.

Government risks subsidising an application arms race

DWP describes its CV builder as helping people create documents suited to online recruitment platforms. In the short term, that may help applicants who would otherwise be disadvantaged by unfamiliarity with recruitment technology.

What helps one applicant becomes less useful when everyone can do it. AI makes it quicker to find vacancies and produce tailored applications. Employers consequently receive more applications that appear to match their requirements. They may respond with tighter filters and additional assessment stages.

Candidates then have a stronger incentive to use AI to overcome those barriers. That risks creating an arms race in which both sides expend more effort without necessarily making better decisions.

Government should be particularly wary of subsidising that cycle. Producing more applications is not the same as improving labour-market mobility. A hundred automatically tailored CVs do not create an additional vacancy or make an unsuitable match productive. Even an increase in interviews may be misleading if employers respond by discounting CV evidence and placing additional hurdles later in the process.

For public services, more digital activity is only useful if it produces better outcomes. Its objective should be to reduce the time, uncertainty and cost involved in matching people with productive work.

Help people prove ability, not manufacture polish

Work Hub could play a larger role than helping people polish CVs. It could help them assemble credible evidence of what they can do, including verified qualifications, employment history, practical work samples and recognised records of skills acquired through work, training, volunteering or caring responsibilities. Someone without a conventional career history may benefit far more from demonstrating competence than from learning how to imitate the language of a job description.

Work Hub could also help people identify which skills a vacancy genuinely requires, where their evidence is weak and what training or experience would close the gap. Work coaches could spend less time correcting document formats and more time helping claimants turn experience into evidence that employers trust.

Employers are already moving towards skills-based hiring. LinkedIn's guidance, for example, recommends structured interviews, hard-skills tests and job simulations that allow candidates to demonstrate how they would perform relevant tasks.

Government should not prescribe one national assessment method. Nor should CVs disappear altogether. They remain a convenient summary and will retain value in many occupations.

Replacing CV screening with opaque personality tests or automated candidate scores would not constitute progress either. Assessments should be relevant to the work, accessible to different candidates and capable of being questioned by a human being. The goal is better evidence without creating a different technological gatekeeper.

Government should improve the signal, not choose the candidate

Labour markets depend on trusted information. Employers need credible signals of capability while workers need a fair opportunity to demonstrate what they can contribute.

When those signals deteriorate, hiring becomes slower and more expensive. Suitable applicants are overlooked, vacancies remain unfilled and recruiters spend more time verifying claims that the application process was supposed to establish.

Government does not need to construct a central hiring algorithm. It can help the labour market generate better information.

DWP can work with employers, sector bodies, colleges and training providers to make qualifications and skills more portable and verifiable. It can help jobseekers understand what employers actually need and build evidence that survives contact with an interview, practical task or workplace. Work Hub remains a worthwhile experiment, but ministers should reconsider what it is meant to optimise before celebrating how many people have used it.

If it becomes a means of helping people identify, develop and prove valuable skills, it could improve both employment support and labour-market productivity.

If it merely becomes a publicly funded factory for job-description-perfect CVs, it will be automating a recruitment model that AI is already making less reliable.

AI has made professional writing abundant. The next challenge for employment policy is making genuine ability easier to see.