Recruitment has always been a race: to find, qualify, and submit the best candidates before the competition does. But over the past 18 months, a new obstacle has emerged - AI-generated CVs that flood the hiring process with misleading or fake information.
This isn’t a niche concern. According to The New York Times' Sarah Kessler, who recently wrote about the phenomenon, the market is now grappling with what she calls “AI sludge” - a wave of mass-generated applications that bury recruiters in low-quality, auto-filled, and frequently inaccurate content.
“Employers are buried in AI-generated résumés,” she writes. “The applications are often misleading, and they’re multiplying at scale.”
– Sarah Kessler, The New York Times (June 2025)
As CEO of a company that processes over 200,000 CVs per month, I can tell you: she’s right.
What Is AI Sludge?
AI sludge refers to the overproduction of CVs generated or heavily “optimized” using generative AI tools like ChatGPT or job board automation scripts. These CVs:
- Use job description keywords to pass screening filters.
- Include exaggerated or fabricated experience.
- Present inconsistencies between job history, qualifications, and public records.
- Bypass early-stage interviews using voice bots or avatar-based AI.
This problem has been compounded by the rise of remote work and global hiring. Now, anyone can apply to any job, from anywhere, and they often do - in bulk. In many cases, candidates submithundreds of auto-generated applications a day.
The result?
Recruiters and staffing firms are overwhelmed, and application quality has plummeted.
Why Traditional Hiring Systems Can’t Cope
The average Applicant Tracking System (ATS) or CRM wasn’t designed to handle this level of noise. Most parsing engines operate with 50–60% accuracy, and they lack the contextual intelligence to flag red flags like:
- Job titles that don’t align with listed companies.
- Skill claims unsupported by any project or experience.
- Gaps in work history glossed over by generic descriptions.
Even diligent recruiters can’t catch everything - especially when they’re reviewing hundreds of applications a day.
In fact, many of the world’s largest firms now report that up to 70% of received CVs contain misleading or unverifiable data.
The Business Impact
Let’s be clear: AI sludge is not just an admin issue. It’s a strategic risk.
- Wasted time: Recruiters spend hours reviewing irrelevant or fake applications.
- Slower time-to-market: Top candidates are delayed behind AI noise.
- Damaged brand reputation: Poor candidate experience and bad hires affect client trust.
- Compliance risks: Inaccurate CVs can lead to failed placements, client disputes, and data integrity issues.
In high-volume hiring environments, this inefficiency directly affects revenue. For staffing firms and in-house TA teams alike, speed and precision are everything.
How to Fight AI With AI—The Right Way
At Allsorter, we’ve built AI to counteract the very problem AI created.
Here’s how modern, trustworthy CV tech needs to respond:
✅ High-accuracy data extraction
Our platform achieves 99%+ parsing accuracy, extracting all relevant data from PDFs, LinkedIn profiles, job board templates, and more.
✅ Trust scoring and validation
We assign a trust score to every CV by cross-checking:
- Job titles and companies
- Qualifications and public data
- Experience claims versus work history
If something doesn’t align, we flag it—immediately.
✅ Profile enrichment and redaction
Our system automatically builds anonymised profiles, branded submissions, and customised CVs matched to job specs—saving hours per submission and removing guesswork.
✅ Real-time performance at scale
We process millions of resumes annually across global clients like Randstad, Adecco, and Manpower—plus enterprise-level integrations with CRMs handling 200 million CVs across 6 million roles.
✅ Risk free user guarantee
Unlike many tools built on third-party LLMs like OpenAI or Gemini, Allsorter uses no external AI models for processing. Instead, we’ve developed our own proprietary Vertical Language Model (VLM) - purpose-built for the HR and recruitment industry. This means no exposure of personally identifiable information (PII), no third-party data sharing, and zero data breaches to date. It's AI built for trust, not trade-offs.
What Recruiters Need to Do Now
Recruitment leaders can no longer rely on old methods. To stay competitive, compliant, and credible in a world of AI sludge, hiring teams must:
- Invest in smarter tools: Point solutions that focus only on formatting or keyword scoring won’t cut it. You need deep, real-time intelligence.
- Build trust into the workflow: A recruiter should never have to guess if a CV is real. Trust scoring and validation should be part of every submission.
- Automate what doesn’t require judgment: Let AI handle the structure. Let humans handle the relationships.
The Future: More Human, Not Less
Despite the problems AI has caused in hiring, I remain optimistic.
As automation increases, so does the value of human oversight. The future isn’t bots replacing recruiters - it’s AI doing the grunt work, so recruiters can focus on what really matters: assessing fit, building trust, and closing great candidates faster.
In this next chapter of recruitment, quality will beat quantity - and the firms who act now to cut through the sludge will win.
The future of recruitment will be human-in-the-loop: AI for speed and consistency, humans for judgment and context. That’s what we’re building at Allsorter.
Because in a world full of AI sludge, trust is your most valuable asset.
Want to learn how we’re helping global firms stay ahead?
Visit allsorter.com or contact us here to explore how trust-first AI can transform your recruitment process by restoring trust,structure and speed!
Listen to the full podcast interview:
👉 Down to Business with Bobby Kerr – Listen here
Let’s connect if you're interested in restoring trust, structure, and speed to your recruitment process. Contact us here.