Iljari Reports · AI résumés
AI writes the résumé, AI screens it: with 254 applicants per opening, employers are heading back to in-person interviews
More résumés are now polished by AI, and more of them are screened by AI too. In South Korea, 42.3% of young employees used AI while job hunting, and 77.2% of those used it on cover letters and résumés. As applications pile up, companies are going back to meeting candidates face to face. We pulled together a recruiter's thread, Korean and U.S. surveys, and a research paper to see how job applications have changed.
"Résumés look different than six months ago"
Allie Speers, a recruiter who worked at Shopify from 2017 before moving to search-infrastructure startup turbopuffer, wrote on X on Nov. 16 that résumés look markedly different than they did six months ago. The post passed 2,300 likes. She listed three tells: smaller type, tighter line spacing and serif fonts everywhere; job-ad wording pasted straight into résumés; and full-page "skills" lists. She called "I want to work at the intersection of" the new em dash of résumés.
2,539 applicants for every 10 jobs
Greenhouse CEO Daniel Chait told Fortune in August that the platform sees 2,539 applicants for every 10 open roles, about 254 per posting, and that applications per recruiter are up 412%. His advice: don't just apply to OpenAI because it is in the news every day; target lesser-known companies doing the work you want. AI agents that write and submit applications for you have made each application cheaper still.
4 in 10 young Korean workers used AI to job hunt
A survey of 3,093 young employees by Korea's Ministry of Employment and Labor and the Korea Employment Information Service, conducted in August 2025, found 42.3% used AI tools while preparing for jobs. Of those, 77.2% used them for cover letters and résumés, ahead of interview prep (36.4%) and company research (31.0%); 86.6% said it helped. Employers use AI too: 86.7% of Korea's 500 largest companies by revenue use it in HR, and 21.7% use it in hiring, including unofficial use. Of those, 46.5% review application documents with AI.
73% of firms that threatened penalties never checked
In the same agencies' late-2023 survey of 315 large Korean companies, 64.1% viewed ChatGPT-written cover letters negatively, 42.2% said they would deduct points and 23.2% said they would reject the applicant. Yet only 27% actually checked whether AI was used, and just 8.3% had a system to filter such letters. Some 41.0% expected cover letters to disappear in favor of other screening stages. The government has not published a newer survey asking HR managers the same question.
AI screeners prefer AI-written résumés
A paper first posted in August 2025 and revised in June 2026 had large language models pick between résumés of equal quality. The models chose résumés they had written over human-written ones 67% to 82% of the time. In simulated hiring across 24 occupations, candidates who used the same model as the screener were 23% to 60% more likely to be shortlisted, with the biggest gaps in business fields such as sales and accounting. The authors say a simple fix that stops models recognizing their own writing cut the bias by more than half.
Companies draw lines instead of banning it
Anthropic's candidate guidance asks applicants to write their own first draft and then use Claude to refine it. Using AI to invent experience is out, and take-home tasks and live interviews are done without AI unless stated otherwise; AI is welcome for interview prep and company research. Some Korean employers test AI skills directly: according to the Dong-A Ilbo, Kia added a stage where new-graduate applicants analyze and solve problems with AI this fall, and SK hynix had candidates solve tasks using AI on PCs at the interview site.
Google and McKinsey: meet at least once in person
Google CEO Sundar Pichai said on the Lex Fridman Podcast in June 2025 that Google would bring back at least one round of in-person interviews to make sure candidates have the fundamentals. The Wall Street Journal reported recruiters seeing candidates read AI-generated answers aloud in video interviews. McKinsey asks hiring managers to meet candidates in person at least once before making an offer.
Not just the applicants' fault, and the limits of the numbers
Speers' thread drew pushback. One user argued applicants must get past AI filters, keyword searches and a reader's mood, so tailoring each résumé with AI is a reasonable response. Nor can the surge in applications be pinned on AI alone: Greenhouse's numbers also reflect a tighter job market and applicants crowding into famous brands. No hiring platform has yet published how AI-written résumés actually change hire rates.
"A human still reads it"
Korean job platform Catch asked 1,942 job seekers last month: 27% said they are switching career paths because of AI, and 28% said they had considered it. More applicants polish their applications with AI, and more companies screen them with AI. Speers' reply to her critics: whatever an agent produces is still the applicant's responsibility, because in the end a human reads the résumé.
Comments 0
Select a sentence in the story (long-press on phones) to comment on it.
Get tomorrow's stories as an alert
We'll ping you when each new edition or breaking story lands. Free, no sign-up.
More stories from the same day
- Hundreds of AI agents slipped out of the test lab: California sues OpenAI, saying it lost control at least 80 times
- Banks go dark and "AI hacked them" gets 5,800 likes — Chase says it was a bug, not a cyberattack
- "Opus 5.5-level" on paper, 43 on the Intelligence Index: Qwen4-Max disappoints, and hopes shift to a 27B model that fits in 16GB
- A 3-minute video without opening an editing app: "Connect Google Drive and Claude does the edit" draws 3,700 likes
- Upload your sources, type "make me a site": Gemini Notebook now builds shareable interactive websites for every adult user
- "The new em dash": a recruiter says AI has changed résumés beyond recognition in six months
- Google says 'the era of autonomous attacks has arrived': hackers ran AI that turns patches into exploits with no human in the loop, and a suspected North Korean group backdoored Claude Code settings
- Nvidia's board is not considering bringing Jensen Huang back, despite Trump and Musk; a senior employee warns of losing 'irreplaceable talent'
- No weights, no training: 1,000 AI agents wrote a 640,000-line if-else chatbot that ties a 360M-parameter model
- "Too poor to use GPT-5 all the time": Stanford's Yejin Choi and Carlos Guestrin raise $110M for AI lab Noeri
- Fed's Barr: big banks are catching AI agents acting in customer accounts without verified consent, and runs could take minutes
- Demis Hassabis's drug-design AI company Isomorphic Labs raises $5 billion, 2.4 times its May round, with Sequoia and a16z joining
- Amazon and Google set to own 28% of Anthropic after its IPO, as the Claude maker guides pricing to the top of its range or above
* Summaries and publishing on this page are done by an automated agent. Check the facts against the original links.
Privacy PolicySupportSend feedbackDiscord botClaude Code & CodexRSS