Press kit
For journalists, podcast bookers, and partner organizations covering AI-detection bias, ESL writing, and academic-integrity due process. Direct contact, founder bio, logo files, and source-cited claims below.
About Diglot
Diglot is a bilingual writing tool that defends real writers from AI-detection bias — built for ESL students, grad students, and professionals whose careful, formal English looks like AI to detectors.
Diglot.ai is a writing platform purpose-built for non-native English speakers who face structural bias from AI-detection tools. Founded in 2023 by Alex Zhovnir, an Estonia-based founder who experienced the «flagxiety» problem firsthand, Diglot combines a connected workspace (drafting, translation, grammar review, paraphrasing, originality checks) with a forthcoming Authorship Certificate — a cryptographically-signed, append-only record of the writing process itself. The company\'s position: AI detectors don\'t actually detect AI; they detect linguistic patterns that systematically disadvantage ESL writers. Diglot\'s product makes building authorship evidence mechanical so that real writers can prove what they actually wrote.
Fast facts
- Founded
- 2023
- Founder
- Alex Zhovnir (Estonia OÜ)
- Headquarters
- Estonia (remote)
- Pricing
- $0 free · $19/mo Spark · $29/mo Pro
- Stage
- Pre-beta · beta launch 2026
- Funding
- Bootstrapped / founder-funded
Key claims with sources
Every claim below is anchored in a primary source. We do not invent statistics; if you spot a number on this site without a citation we want to know. Email press@diglot.ai.
- 01
False AI-detection accusations are now in federal court
A California family submitted 1,162 pages of evidence to defend their daughter against a Turnitin 76% AI-generated flag on an English essay. The school district rejected the evidence. The family filed a federal lawsuit (Doe v. PAUSD) in May 2026.
Source: Hoodline, May 11 2026
- 02
Landmark state ruling: AI-detection-based discipline ruled "devoid of reason"
In January 2026, New York State Supreme Court (Nassau County) ruled that Adelphi University's AI-detection-based academic penalty was "without valid basis and devoid of reason." Sanctions were rescinded and the student's record was expunged. The case is considered a precedent for due process in AI disputes.
Source: Justia (Newby v. Adelphi)
- 03
AI detectors are statistically biased against non-native English writers
A 2023 Stanford study found GPT detectors flagged over half (61%) of essays by non-native English speakers as AI-generated. Native-written essays were flagged at a far lower rate. ESL writing — narrower vocabulary, consistent grammar, formal structure — triggers the same statistical signals AI detectors look for in machine text.
Source: Stanford HAI / Patterns research
- 04
Turnitin itself is shifting "from detection to transparency"
In February 2026, Turnitin CEO Chris Caren publicly stated the company is shifting its product strategy "from detection to transparency" — implicit acknowledgment that punitive percentage scores were never designed as infallible verdicts. Universities including Vanderbilt, Waterloo, and MIT have rolled back AI-detection enforcement since 2023.
Source: Turnitin Feb 2026 press release
Hero quotes
Pull-ready quotes from Alex Zhovnir on the Diglot position. Attributable: «Alex Zhovnir, founder of Diglot.ai». For tailored quotes or interviews, email press@diglot.ai.
"Building authorship evidence before an accusation is no longer paranoid. It is the new minimum for any non-native English writer producing high-stakes work."
"The cleaner your textbook English, the more it looks like a model wrote it. That's not a calibration bug — that's the entire mechanism of how AI detectors decide what counts as human."
"We are not anti-AI. We are pro-evidence. The question is who owns the burden of proof when a real student's real essay gets flagged — and right now, the answer is: the student."
Brand assets
Logo files for editorial use. Diglot is a registered word mark (in progress); please do not modify the lockup, alter the lime accent color, or place the logo on backgrounds that compromise legibility.
Founder
Alex Zhovnir
Founder & CEO, Diglot.ai
Alex Zhovnir is the founder of Diglot.ai, a bilingual writing platform built for non-native English speakers. He is non-native English himself (Ukrainian first language), and the «flagxiety» problem is autobiographical — he experienced being penalized for formal English writing that triggered AI-detection signals. Prior to Diglot he worked across international companies where the gap between native and non-native English writers in high-stakes documents was a consistent structural cost. He runs Diglot from Estonia (Diglot OÜ in registration).
Direct contact: alex@diglot.ai · LinkedIn
Topics we can speak to
- The structural bias of current AI-detection tools against ESL writers
- Why «flagxiety» (the constant low-grade fear of being wrongly accused as AI) is becoming a generational issue for non-native English students
- The wave of 2025-2026 legal cases against AI-detection-based academic discipline (Doe v. PAUSD, Newby v. Adelphi, Doe v. Yale)
- The growing list of universities rolling back AI-detection enforcement (Vanderbilt, Waterloo, MIT)
- The technical mechanics of «perplexity» and «burstiness» that make ESL writing statistically indistinguishable from AI output
- The case for cryptographically signed authorship provenance as the answer to the «who wrote this» question
- What writing tools owe their non-native English users in the AI-detector era
