Skip to content

Blog

Why ThaiGraph publishes its sources

Why ThaiGraph publishes its sources

What this post is

ThaiGraph cites every factual claim in its editorial pages, names every author with a real biography page, and builds its directories and galleries from primary work rather than by scraping or aggregating other sites. This post explains why. The short version is that the AI-slop era of design writing has made the basic editorial fundamentals (source, attribution, originality) rare enough to matter again as a quality signal. The longer version follows.

The problem we are writing against

An estimated 74% of web content published in 2025 contained AI-generated components, up from 17% in 2022 (Originality.ai, 2025); Google’s March 2024 Helpful Content update and March 2026 core update specifically targeted scaled AI content lacking original research and named author attribution (Google Search Central, 2024 and 2026). The immediate effect on the reader is that search results for most design queries now return variations of the same aggregated, unsourced, and frequently incorrect AI-generated summary. The practical effect is that the baseline expectation for how design writing should work has deteriorated in measurable ways.

Specifically: unsourced statistics have become the rule rather than the exception; named-author bylines are increasingly absent; the same claim appears identically across dozens of sites with no original source attribution; directories and galleries are aggregated from other sites without permission or verification; claims about design history, practice, and industry are made with a confidence that does not match the underlying reporting.

The editorial principles

ThaiGraph runs on four principles: every claim sourced, every author named, every entry original, and no AI-generated imagery used in place of real work. Each is a concrete rule rather than a stated preference.

  1. Every claim sourced. Every statistical, historical, industrial, or technical claim in a ThaiGraph editorial page links to a named source with a date of access. Where we cannot source a claim to our satisfaction, we mark it explicitly as unsourced and note why in our missing-sources.json audit file (visible to our editorial team and available on request).

  2. Every author named. Every editorial piece is bylined by a real named author with a biography page on this site. The author’s credentials, experience, and areas of expertise are stated on the biography page. Author bylines do not rotate pseudonymously; our authors are real people whose professional work is checkable.

  3. Every entry original. Directory listings, gallery inclusions, designer and studio profiles, font entries — each is built from primary work: public records, studio portfolios, award archives, direct outreach, and submissions. We do not scrape or aggregate other sites. When we draw on published sources (for example the CEA’s Creative Economy Indicators, or ThaiGa’s membership report), we cite them.

  4. No AI-generated imagery in place of real design work. ThaiGraph is about Thai design. The work shown in galleries, profiles, and case studies is real Thai design work, credited to the designer or studio who produced it. We may use AI-generated imagery for abstract editorial illustration (hero images, decorative pattern, stylised site navigation) where it is not standing in for a specific designer’s work and where it is visibly non-photorealistic.

Why we are showing our work

The decision to publish source-citation and editorial policy visibly — rather than treating them as internal editorial practice — reflects a simple position: in a content environment where most claims are unsourced, publishing sources is a quality signal worth making visible, and readers increasingly look for it. Our source citations are rendered at the bottom of every editorial page. Our author bios are linked from every byline. Our editorial policy is this post.

The secondary reason is specific to AI-era search and citation. Large-language-model search and retrieval systems (ChatGPT search, Claude search, Perplexity, Gemini’s search) preferentially cite content that is well-structured, clearly sourced, and attributable to named experts. The alignment between writing for real readers and writing for AI-era discovery is, conveniently, complete. Writing sourced and attributed editorial work is good practice for its own reasons and also performs better in AI citation — a rare alignment worth using.

The alignment with Thai industry norms

Thai design industry norms around AI, sourcing, and originality have been shifting visibly in the last two years in the same direction as ThaiGraph’s editorial policy — 61% of Thai heritage and hospitality clients now specify “no AI-generated imagery” in their design briefs, up from 14% in 2024 (ThaiGraph Client Brief Analysis 2026). The editorial discipline we are asking of ourselves in writing about Thai design is the same discipline Thai clients are now asking of their studios. This is not coincidence — the underlying pressure is the same in both domains. When the baseline reliability of unsourced and unattributed work goes down, named attribution and verified sourcing become quality signals.

For the full picture on the Thai industry’s AI response, see State of AI in Thai Design 2026.

What this costs us

Running editorial this way is slower than the AI-aggregation alternative by a meaningful factor — our estimate is that ThaiGraph’s source-first production produces editorial content at about 30–40% of the per-day word count of equivalent aggregated-and-spun AI content, and at materially higher per-word cost. We think the cost is worth it for three reasons: the work is more useful to real readers, the site’s authority compounds rather than decays over time, and the editorial standard sustains the specific audience we are trying to serve (working Thai designers, international designers working on Thai projects, clients evaluating Thai studios).

We are not an AI-resistant publication — our authors use AI tools in research, drafting, and editing in the same ways most professional writers now do. The discipline is at the editorial output rather than the production workflow: what we publish is sourced, named, and verifiable, regardless of what tools were used in its production. That boundary — between tool and output — is the practical line we are drawing.

Limitations and where we can improve

ThaiGraph is not perfect at this, and we know where the gaps are. The current limitations:

These are real limitations. We do not pretend otherwise. Where a claim is on shaky ground we mark it and we work on improving the underlying sourcing.

What we ask of readers

If you find a factual error on ThaiGraph, tell us. Contact the editorial team through the site’s contact form. We correct the page, log the correction in our editorial log, and credit the reader if they want to be credited. If you are a Thai designer or studio whose work we have not yet represented, submit it — we publish on an ongoing rolling basis.

The editorial standard exists because of the reader. The reader makes it sustainable.

Go deeper

For what is on the site right now, see Welcome to the new ThaiGraph. For our author biographies, see the about section. For the industry context in which these editorial decisions were made, see the Thai Graphic Design Industry overview and State of AI in Thai Design 2026.

Information verified as of April 2026

Sources

  1. An estimated 74% of web content published in 2025 contained AI-generated components, up from 17% in 2022.Originality.ai — State of AI Content 2025 Report (accessed Apr 9, 2026)
  2. Google's March 2024 Helpful Content update and March 2026 core update specifically targeted scaled AI content lacking original research and named author attribution.Google Search Central — March 2024 Helpful Content Update and March 2026 Core Update Announcements (accessed Apr 9, 2026)
  3. Among Thai heritage and hospitality clients, 61% specify 'no AI-generated imagery' or equivalent language in design briefs as of 2026, up from 14% in 2024.ThaiGraph Client Brief Analysis 2026 — 180 briefs sampled from ThaiGa member studios (accessed Apr 8, 2026)