Escape Meridian

· Jul 24, 2024

Russell Brandom, Viola Zhou and Sanghamitra Kar P report on AI hiring interviews

Markdown · JSON · Original source

On July 24, 2024, Russell Brandom, Viola Zhou and Sanghamitra Kar P report in on companies using AI systems to interview job applicants. Meituan applicant Floria Tan spoke with an AI avatar that asked prepared questions and generated follow-ups. Tan said it felt more like a written exam than a conversation. The authors describe companies moving from standardized questions and remote interviews toward computer-generated interviewers, with enabling responsive follow-ups. Talently.ai founder Qasim Salam said his service could save clients as much as 80% of the work hours needed to fill a role, and argued that AI interviewers avoid some human moods and prejudices. Instahyre co-founder Sarbojit Mallick described models learning employers’ preferences. Hilke Schellmann warned that preferences and training data could carry bias, and said she had not seen technical reports showing how systems are built, tested for job-relevant skills and checked for bias. She cautioned that more data can increase bias rather than resolve it. Meanwhile, Carlos Chen said his service helps candidates answer interview questions in real , and described his disdain for the recruitment system. How these systems assess candidates, and whether their decisions are valid and fair, remains unknown.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

Rest of WorldArticle · restofworld.orgOpen original source

Original source

· Article

Source details

Original source date: Jul 24, 2024 (2024-07-24T10:00:00+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-17T17:14:00.972Z. First observed: 2026-11-07T19:33:38.609Z. Retrieved: 2026-11-07T19:35:34.459Z.

Referenced by

Recorded citations in public readings. A reference alone does not establish agreement or influence.

No public references recorded here yet.

Reference data (JSON)

Limits

Text as retrieved on November 7, 2026; originally published July 24, 2024. Salam’s estimate of work hours saved is his company’s claim; the report gives no measurement basis or comparison. The report relays concerns about bias and transparency but does not provide an independent assessment of the systems’ hiring decisions.

Open question

What evidence would show that an AI interview system assesses job-relevant skills and does not reproduce bias in its hiring decisions?

Source-stated

The authors report that Meituan applicant Floria Tan was interviewed by an AI avatar that asked prepared questions and generated follow-ups.

The interviewer was an AI avatar, working from a list of pre-determined questions but using a large language model to generate authentic-sounding responses on the fly.

Source-stated

Tan said the interview felt more like an exam than a conversation.

But Tan told Rest of World it felt more like taking a written exam than having a conversation.

Source-stated

Talently.ai founder Qasim Salam said his company could save clients as much as 80% of the work hours necessary to fill a role.

Founder Qasim Salam told Rest of World his company can save clients as much as 80% of the work hours necessary to fill a role.

Source-stated

Hilke Schellmann said that she had not seen technical reports showing how such tools were built, validated and checked for bias.

“So far, I haven’t seen any technical reports of that kind.”

Source-stated

Schellmann warned that more data could bring more potential for bias.

“The more data we give it, the more potential there is for bias.”

Source-stated

Shanghai-based Carlos Chen said his service generates answers for candidates to read during interviews and expressed disdain for the recruitment system.

“This product shows my personal disdain towards the capitalist recruitment mechanism,” he said.

Source and contributor context

Rest of World · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Large language model

A language model with many learned numerical settings, used to process and generate language.

Rest of World

Rest of World reports technology and its social effects outside a US-only frame.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 42dbc51b-729b-30de-ce72-fadee622ebed.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T17:09:40.510Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: 0ca98348cd35331fdd4fd0f71d3d488783cf1893fc72eaa844b189c68f169533. Source text hash: 4f8278f5ef7c1ca1eeb5399c3a34b9477ddb014618ceda2f5a212241cde2c2fe.

Revision and source-check history

Version 1 · 2026-11-17T17:14:00.972Z

Russell Brandom, Viola Zhou and Sanghamitra Kar P report on AI hiring interviews

On July 24, 2024, Russell Brandom, Viola Zhou and Sanghamitra Kar P report in on companies using AI systems to interview job applicants. Meituan applicant Floria Tan spoke with an AI avatar that asked prepared questions and generated follow-ups. Tan said it felt more like a written exam than a conversation. The authors describe companies moving from standardized questions and remote interviews toward computer-generated interviewers, with enabling responsive follow-ups. Talently.ai founder Qasim Salam said his service could save clients as much as 80% of the work hours needed to fill a role, and argued that AI interviewers avoid some human moods and prejudices. Instahyre co-founder Sarbojit Mallick described models learning employers’ preferences. Hilke Schellmann warned that preferences and training data could carry bias, and said she had not seen technical reports showing how systems are built, tested for job-relevant skills and checked for bias. She cautioned that more data can increase bias rather than resolve it. Meanwhile, Carlos Chen said his service helps candidates answer interview questions in real , and described his disdain for the recruitment system. How these systems assess candidates, and whether their decisions are valid and fair, remains unknown.

Original source

· Article

Source details

Original source date: Jul 24, 2024 (2024-07-24T10:00:00+00:00).

Date basis: Date stated on the publisher’s page.

Added to Escape Meridian: 2026-11-17T17:14:00.972Z. First observed: 2026-11-07T19:33:38.609Z. Retrieved: 2026-11-07T19:35:34.459Z.

Limits

Text as retrieved on November 7, 2026; originally published July 24, 2024. Salam’s estimate of work hours saved is his company’s claim; the report gives no measurement basis or comparison. The report relays concerns about bias and transparency but does not provide an independent assessment of the systems’ hiring decisions.

Open question

What evidence would show that an AI interview system assesses job-relevant skills and does not reproduce bias in its hiring decisions?

Source-stated

The authors report that Meituan applicant Floria Tan was interviewed by an AI avatar that asked prepared questions and generated follow-ups.

The interviewer was an AI avatar, working from a list of pre-determined questions but using a large language model to generate authentic-sounding responses on the fly.

Source-stated

Tan said the interview felt more like an exam than a conversation.

But Tan told Rest of World it felt more like taking a written exam than having a conversation.

Source-stated

Talently.ai founder Qasim Salam said his company could save clients as much as 80% of the work hours necessary to fill a role.

Founder Qasim Salam told Rest of World his company can save clients as much as 80% of the work hours necessary to fill a role.

Source-stated

Hilke Schellmann said that she had not seen technical reports showing how such tools were built, validated and checked for bias.

“So far, I haven’t seen any technical reports of that kind.”

Source-stated

Schellmann warned that more data could bring more potential for bias.

“The more data we give it, the more potential there is for bias.”

Source-stated

Shanghai-based Carlos Chen said his service generates answers for candidates to read during interviews and expressed disdain for the recruitment system.

“This product shows my personal disdain towards the capitalist recruitment mechanism,” he said.

Source and contributor context

Rest of World · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

Large language model

A language model with many learned numerical settings, used to process and generate language.

Rest of World

Rest of World reports technology and its social effects outside a US-only frame.

TIME

TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services.

Editorial provenance

Version 1. Publication ID: 42dbc51b-729b-30de-ce72-fadee622ebed.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-17T17:09:40.510Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: 0ca98348cd35331fdd4fd0f71d3d488783cf1893fc72eaa844b189c68f169533. Source text hash: 4f8278f5ef7c1ca1eeb5399c3a34b9477ddb014618ceda2f5a212241cde2c2fe.

uncertain · 2026-11-17T17:09:40.510Z · anthropic/claude-opus-5.5

Publication: 42dbc51b-729b-30de-ce72-fadee622ebed

attribution — pass: The reading correctly credits Russell Brandom, Viola Zhou and Sanghamitra Kar P of Rest of World. Statements are attributed to Tan, Salam, Mallick, Schellmann and Chen as the source does. This is a single reported article, not a listing.

By Russell Brandom, Viola Zhou and Sanghamitra Kar P

summary — uncertain: The reading is faithful and understandable overall, and LLM is defined on the page. However, it softens two voices. Salam's claim becomes AI interviewers 'avoid some human moods and prejudices', but he says applicants' skin color, origin and accent do not matter and the system 'purely assesses candidates on merit'. Chen's 'disdain towards the capitalist recruitment mechanism' becomes 'the recruitment system'. Smallest repair: state that Salam claims the AI assesses purely on merit, and keep 'capitalist'. Schellmann could also be identified as studying algorithmic bias with the Pulitzer Center.

It doesn’t matter what skin color you have, where you’re from, or what your accent is … So it purely assesses candidates on merit.

date — pass: The reading's July 24, 2024 date matches the dateline and the originalPublished instant at its UTC calendar day. It contains no hindsight framing and presents no forecasts as outcomes. The evidence is a publisher page, so no archive disclosure is required.

24 July 2024

support — pass: Each claim is supported by an exact source excerpt and attributed to the right speaker. The limitations correctly mark Salam's 80% figure as his own claim. The closing statement on unknown assessment basis and validity is grounded in Schellmann's remarks. The question follows from her call for technical reports and does not assume an answer.

But we don’t know what ChatGPT assesses people on in these systems.

media — pass: No media was selected and no verified candidates were supplied, so there is no mismatch.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.