Why CV-slop has become the biggest AI-related issue in high-volume recruitment

Over the last few years, we have noticed a big shift in applications per job posting, with an average growth of 2.6-3x since 2024. This means that a job posting which once used to get 200 applications now gets 600 - and this is cemented by the numbers we see in reports. Over 12 million job applications are submitted on LinkedIn every day in 2025, or over 9,000 per minute, creating enormous top-of-funnel volume at companies that rely on LinkedIn as their primary sourcing channel.

 

This isn’t a phenomenon unique to any one region either. High-volume hiring pressure is mounting across APAC, the UK, the USA and Europe too, particularly in sectors like retail, health and logistics, where seasonal and large-scale hiring make the applications-per-posting problem even more acute.

 

This situation has created issues for both ends of the coin - the candidates and the companies hiring them. According to research, 75% of CVs are never seen by a human; they are filtered out by the company’s Application Tracking System (ATS). Back in 2019, 88% of employers believed ATS systems could screen out high-quality candidates, but that landscape is now shifting too, thanks to over-automation.

 

Google’s DeepMind was recently in the news for using manual forms alongside its AI screenings because it trusts that its AI might be screening out deserving candidates. This raises eyebrows: if every candidate has access to myriad AI tools and companies hone their best tech stack, why does time to hire still take 30-45 days?

 

 

The rise of ’keyword vomit’

From a candidate perspective, it’s now easy to copy and paste a job description into an AI tool, extract keywords and add them to a CV. More and more people are giving ATS tools exactly what they want to hear to get past the barrier, which many people weren’t aware of earlier.

 

This has bombarded talent acquisition teams with identical CVs, making it difficult to sift through 600 applications that seem ’perfect’. This has backfired a little because a tool initially made to cut through the noise is now being surpassed by AI-savvy candidates.

 

 

The general consensus on AI interviews

AI adoption is outpacing trust. Nine out of ten companies use AI in their recruitment process. The efficiencies are undeniable: AI interviews 40 to 80 people daily, compared with the six to ten people interviewed manually. Structured AI interviews also outperform unstructured human interviews in predicting potential employees’ performance. However, this does not seem to impress the candidates, as two-thirds of them have been through AI interviews but only a quarter of them trust it to judge them fairly, reporting almost similar levels of bias in AI interviews compared to human interviews, with a fear that algorithms unfairly discriminate against neurodivergent applicants, non-native speakers, and non-traditional career paths.

 

In 2026, consensus is settling on a hybrid approach. AI is assigned the repetitive 70% of the work, while 30% is reserved for the human components of emotional intelligence and relationship-building.

 

 

Transparency in high-volume recruitment

Giving individual feedback and updates was easier when the candidate pool was scarce. Now, 1 in 2 candidates claim that they are being ghosted by the employer. No update, no explanations. Most get an auto-rejection email a minute after they apply. From a candidate’s point of view, this stains the employer’s brand, with direct repercussions for the employer’s balance sheet. A strong employer brand can cut cost-per-hire by up to 50% and reduce turnover by 28%. Companies that invest deliberately in employer branding see roughly a 3.5x return over three years, measured across faster fulfilment, lower hiring costs and higher retention.

 

This can be avoided by making your ’sources of truth’ known - especially in high-volume situations where individual feedback cannot be provided. Every candidate deserves to know where they may have slipped through the cracks. In our experience, a best-in-class screening system should always match skills listed against the candidate’s experience. Slightly controversial, but it’s also important to verify the candidate’s presence across public profiles like LinkedIn. While this might come off as a little invasive, more and more companies are mapping candidates’ social profiles to check for genuineness and role or culture fit.

 

A problem created by AI cannot be solved by adding more AI layers; screening processes now need greater transparency around false-negative rates and a sense of ownership from employers when AI tools hallucinate or don’t deliver optimal results.

 

 

The line between efficiency and over-automation

As more companies adopt AI, it has become imperative to realise that speed without judgement is just noise at scale. Efficiency and automation cater to different questions - while efficiency asks "how do we get this done quickly?", over-automation forgets to ask "should a person be making this decision instead?" The moment a person’s career becomes a number in a spreadsheet, a line is crossed. In high-volume scenarios, recruitment takes the form of a funnel, but it’s still a relationship to be earned. AI should do the heavy lifting so a human can carry the judgement, because the best hiring decisions will always need a person at the heart of the recruitment process.

 

This is what the hybrid ratio looks like in practice: a multinational conglomerate recently compressed a six-month hiring cycle to just two weeks. A 300-role campaign generated tens of thousands of candidates in a single requisition cycle. AI screened, ranked and interviewed the whole pool on top of their existing ATS, while recruiters kept control of the workflow they already swore by. When you adopt an optimal human-to-tech ratio, a high-volume recruitment funnel is bound to deliver exceptional results.

 

---

 

Presha Chokshi is Research and Marketing Manager at Careerfit

 

Main image courtesy of iStockPhoto.com and AndreyPopov