HumanizeMyAI in the press.
Independent coverage of the tool, its founder, and the computational-linguistics research behind it. We link out to each piece in full rather than reprinting it, so you read it where it was published.
Last reviewed September 13, 2026
Covered by
- TechRadar Pro
- Yahoo Tech
- MSN
- Nate News
- ValiantCEO
- inkl
- DigitalToday
- Telset.id
- GoodFirms
- TechRound
- Leaders Perception
- Alanya Kent Konseyi
- Gazete Alanya
- Son Alanya
- Alanya Postası
- Haberi Bizden
- Mega TV
Which coverage should a journalist read first?
These four pieces are the ones we would hand a journalist first. TechRadar Pro broke the 18,989-abstract study on 1 September 2026, Yahoo Tech carried it the same evening and Nate News put it in front of a Korean audience the next day. The fourth is a founder interview, for the company rather than the research.
AI may be changing the meaning of human writing as rewritten machine text quietly enters the scientific record
“AI only needs its rewritten output to be accepted once as human. After that, the disguise becomes part of the answer key.”
AI vocabulary is creeping into research written under human names
Efosa Udinmwen carried the study to Yahoo's technology audience the same evening.
AI vocabulary is creeping into research written under human names
Efosa Udinmwen ran the TechRadar Pro report on Microsoft's news front page.
AI가 자주 쓰던 단어, 인간도 쓰기 시작…탐지기 혼란
put the finding on one of Korea's largest news portals.
Fırat Mıhcı of HumanizeMy.ai: Transforming AI Solutions as a Solo Founder
“AI writing technology should be evaluated through evidence, not marketing language.”
Where has the founder been interviewed?
Fırat Mıhcı has given three on-record interviews: two with ValiantCEO, in September 2026 on pricing by workload and in August 2026 on running the company as a solo founder, and one with Leaders Perception on preserving meaning over detector scores. Each card links out to the full conversation on the publisher’s own site.
A conversation with founder Fırat Mıhcı HumanizeMy.ai is a research partner of this survey of 192 marketers, which quotes our founder on why research only earns trust when anyone can check it.
Read the full interview at GoodFirms“At HumanizeMy.ai we publish our detector's error rate on the product page: it wrongly flags genuine human writing under 0.2% of the time.”
By Gina Marrs. A conversation with founder Fırat Mıhcı on why publishing numbers anyone can test is how a small AI company answers the new calls for AI oversight.
Read the full interview at TechRound“Do not make customers upgrade to trust the product; make them upgrade because they need more of a product they already trust.”
A conversation with founder Fırat Mıhcı on pricing by workload instead of locking features away, so every paid plan gets the same product.
Read the full interview at ValiantCEO“AI writing technology should be evaluated through evidence, not marketing language.”
A conversation with founder Fırat Mıhcı on running the company as a solo founder and holding the product to published evidence.
Read the full interview at ValiantCEO“If the meaning changes, the metric is irrelevant.”
By Eric Zavala. A conversation with founder Fırat Mıhcı on why the tool optimises for preserved meaning rather than a detector score.
Read the full interview at Leaders PerceptionWhat has the press written about the research?
In September 2026 the technology press picked up our study of 18,989 arXiv abstracts, which tracked AI-associated vocabulary rising roughly sevenfold in computational linguistics after ChatGPT while a mathematics control stayed flat. That contrast is the evidence that the words are spreading as style rather than as subject matter, which is precisely why word-list detection keeps misfiring on real writers. It ran in English, Korean, Japanese, Chinese, Vietnamese and Indonesian inside two days. The lab has since run the same kind of word count on diplomacy, first in a 2015 to 2025 study of UN General Assembly speeches and then in a follow-up on the 2026 General Debate.
One finding, six languages, inside two days
What they reported
AI-associated vocabulary rose roughly sevenfold in computational linguistics after ChatGPT, while a mathematics control stayed flat.
1 September 2026
The study breaks
- TechRadar Pro · first report
- Yahoo Tech · the same evening
2 September 2026
Five more languages in a day
- Nate News · Korean news portal
- DigitalToday · Korean, English, Japanese, Chinese, Vietnamese
- Readers.id · Indonesian
Later in September 2026
Dated by month only
- MSN · ran the TechRadar Pro report
- inkl
- Telset.id · Indonesian
English Korean Japanese Chinese Vietnamese Indonesian
The disguise becomes part of the answer key: researchers find that AI is redefining what human writing means
AI가 자주 쓰던 단어, 인간도 쓰기 시작…탐지기 혼란
AI-favoured words spread to humans, confusing detectors
AIが好む語彙が人間の文章に広がり、検出ツールを惑わせる
AI 常用词汇渗入人类写作,检测工具受到干扰
Từ vựng mang dấu ấn AI xuất hiện trong bài viết của con người, làm giảm độ tin cậy của công cụ nhận diện
Studi: Kosakata AI Makin Sulit Dibedakan dari Tulisan Manusia
What was the Alanya City Council study?
Independent Turkish media covered our published study of 12,222 visitor reviews, an NLP analysis run for the Alanya City Council. The tool and the study come from the same lab; reading human language carefully, at scale, is the skill underneath both. Full methodology and open data live on the research hub, and the study page itself explains how the work with the Council came about.


Writing about HumanizeMyAI or the research? Reach Fırat directly. The studies and their data are public, and the evidence protocol behind every published claim is documented.