Version 1.8, run#10, of MECILDI has been reached, with errors corrections, improvements in error management and for the last version a functional enhancement with the measurement for each languages for the percentages of sites having a linguistic version in that language (the sum of those percentages is equal t 100% multiplied by the rate of multilingualism, as we do manage multilingual web sites.

The MECILDI page has been updated as well as the state of multilingualism page and the public files of all the results have been both updated and complemented with all relevant intermediary studies results. The Excel file holds now 10 sheets, the last and final results named FINAL and complementary ones as described below:

> V1.7 (run 8) results on Tranco collected June 12th 2026
> V1 (first run) results on Tranco collected Nov. 2025, with confidence interval for each field and comparisons with W3Techs and OBDILCI model
> Sensibility study for factor alpha (extrapolation) and Accept-Language Header.
> V1 (run3) on Tranco collected April 2026, comparison with run1
> V1 (run4) same Tranco from April 2026 but different random sample, control of statistical consistence
> Study of trends using the facility of Tranco to provide older lists, back to 2019
> Comparison between OBDILCI model for the whole www results and Tranco results for the million most visited web sites
> Result using only Crux data instead of Tranco with parameters set to exclude sub-domains and repetition of sites for the same organization (run9)
> Focus on the sites with Hreflang= instruction and capture of their specific data

The post PROJECT MECILDI V1 CONCLUDED appeared first on OBDILCI.

AN INTERESTING AI EXPERIMENT POST MECILDI

After the announcement of MECILDI and its preprint publication, OBDILCI conducted an experiment with eleven AI applications. Several objectives were drawn for that experiment:

1. Check in what proportion the progresses realized by OBDILCI in the theme of measuring languages presence online are reflected in AI’s answers.
2. Assess the potential impact of MECILDI in that field by asking IAs to react to the preprint and its potential impact.
3. Try to “educate” scientifically  the AIs in a field dominated by an extremely popular, yet strongly biased data provider.
4. Assess the possibility that the improvement made by AIs in a conversation could impact future conversations with other users.
5. Assess the impact of the conversation language in the answers.

The results of that experiment are documented here with links to every specific conversation.

There is a large level of convergence in the answers of the AIs, within a split in 3 categories according to the fact that OBDILCI’s data could be mentioned in the first answer, the second one or in the third one, only after explicit mention.

The choice of language has no effect in that experiment except in two AIs. One case, DeepSeek, showing a tremendous impact in what could be called a macro hallucination triggered by a massive bias.

THE CONCLUSIONS

1. The progresses are well reflected, in half of the case from scratch, however in 25% of the AIs, not until the reference to OBDILCI is explicitly made.
2. There is a consensus that MECILDI is a game changer, one AI even call it a chess mate.
3. The education process works fine by opposing peer reviewed data to undocumented one.
4. All the AIs but one pretended they will answer totally differently to the same question starting immediately. Test proved this is false. There are still some gray areas in this question, which appears as a crucial issue for the future of AI.
5. Can we reason with an AI using solid arguments and make it change a view based on popular consensus to one based on sound scientific principles, even if less popular? The answer is yes, without a doubt, and thankfully so. This is why using AI correctly requires a good dose of critical thinking.
6. The change of language did not modify the answers, except for CoPilot and DeepSeek. The singular case of DeepSeek, and the massive anti-Francophonie bias which emerged, motivates the following additional questions, some remain unanswered.
7. Can the creators of an AI intentionally bias its responses in a specific direction? Yes, by feeding it with data that heavily contain this bias.
8. Do the Chinese authorities have a well-established anti-Francophone stance?
9. Did this anti-Francophonie bias in DeepSeek emerged in that experience as a hallucinatory accident, or is it “programmed” to express itself systematically?

The post AN INTERESTING AI EXPERIMENT POST MECILDI appeared first on OBDILCI.

The pre-recorded version of OBDILCI’s presentation at the 7th European Conference on Multilingualism, May 22th 2026, in the session “Multilingualism and Societal and Cultural Challenges” – TB 10 – Multilingualism, a major issue for the Internet and AI, on the theme “Languages ​​and Multilingualism on the Internet: State of the Art and Perspectives” – VIDEO (in French, 20mn)

The post ASSISES DU PLURILINGUISME appeared first on OBDILCI.

Every thing you would like to know about the new OBDILCI program MECILDI to measure languages and multilingualism in any series of web site and its first outcomes is now documented in the following preprint in ResearchGate: MECILDI: computing languages and multilingualism indicators of any series of web sites with due attention to web multilingualism – Application on TRANCO and comparation with W3Techs, Daniel Pimienta and Mike Field, March 2026

The post PREPRINT ON MECILDI appeared first on OBDILCI.

Study on multilingualism of gTLDs of France

As second part of the project funded by DGLFLF, MECILDI was applied in the following geographic/linguistic gTLDs with the goal to collect indicators of multilingualism and language’s proportion:  .alsace, .bzh (Bretagne), .corsica, .eus (basque), .gp (Guadeloupe), .nc (Nouvelle Calédonie) and also .quebec and .paris.

Some 20 different indicators were collected such as percentage of multilingual web, rate of multilingualism, average number of languages per multilingual site, percentage of IDN use, different types of errors, etc.

The post Study on multilingualism of gTLDs of France appeared first on OBDILCI.

The linguistic digital divide

The French-language Canadian journal Minorités linguistiques et société publishes in its Issue 23, 2024, dedicated to “Digital Technology and Linguistic Communities in Minority Settings: Roles, Impacts, and Challenges,” an open-access article by Virginie Hébert and Maria Bodron titled “The Linguistic Digital Divide: A Framing Analysis of the Literature on Language Inequalities in the Digital Environment.” Link: https://doi.org/10.7202/1124262ar

This article is interesting for several reasons:

• To our knowledge, this is the first time that a systematic effort to research, compile, and analyze has been undertaken at the intersection of the digital divide and linguistic diversity. The dominant view of the digital divide—one that is entirely technology-focused and obsessed with access—has historically overshadowed fundamental issues such as digital literacy and linguistic diversity.
• For their systematic review, the authors cross-referenced three categories, each comprising about ten terms, synonyms for divide, or for digital or linguistic, and focused their search for sources on both academic literature and gray literature (which we consider very wise for a topic where civil society has been highly active in the field of action research, and often ahead of the academic world). They systematized their search, covering the period 1995–2024, using the databases Érudit, CAIRN, ERIC, HAL, Scopus, and Google Scholar, then supplemented this with manual searches within international organizations competent in these areas (such as UNESCO), and subsequently used the bibliographies of the identified sources to systematically expand the number of sources to be analyzed.
• The methodological approach was followed by a method known as frameworking, which thus employs another form of systematization, this time focused on content analysis.
• Last but not least, the result of such a thorough examination of all the concepts underlying the “linguistic digital divide” could have turned out to be unappealing, hard to digest, or even unpalatable—but none of that is the case! The result is a very enjoyable read, as well as an enriching one, connecting and fostering dialogue between sources that might otherwise remain unaware of one another, as bridges between gray literature and academic scholarship are rare.

In short, this is a reference work.

This is an area where AI is making a major impact, which should encourage the authors to update the text in a few years—or sooner—to focus on the effects of AI on the subject.

The post The linguistic digital divide appeared first on OBDILCI.

MECILDI V1 FIRST RUN ON TRANCO SERIES

April 2026: This is an historical moment for whoever is interested in languages@internet. For the first time, a language detection program is designed with due consideration to web multilingualism, correcting the strong bias of previous methods favoring English.

Here are  the first  results of MECILDI version 1, applied to the Tranco sampling of one million  web sites the most visited.

Guess what? All of our main documented predictions are confirmed!

The percentage of web pages in English in TRANCO series is 22%

The post MECILDI V1 FIRST RUN ON TRANCO SERIES appeared first on OBDILCI.

Human Resilience for the AI Age

An essay, co-written with Luis Germán Rodríguez Leal, is published in the last production of the always impacting and stimulating Elon University’ Imagining the Digital Future Center :  “Building a Human Resilience Infrastructure for the AI Age”.

This essay focuses the urgency and importance of pervasive and comprehensive Digital and Information Literacy, as a necessary strategy to face AI challenges and draws 3 possible future scenarios, Utopia, Status Quo or Dystopia.

The post Human Resilience for the AI Age appeared first on OBDILCI.

Le français dans le monde

Publication by OIF/Gallimard of the report about French language in the world in which we have contributed to the chapter about languages in the Internet, pp 187-193.

The full report in French is accessible here (42MB).

The post Le français dans le monde appeared first on OBDILCI.

A new section have been added under the title MAIN PROJECT 2: MECILDI. MECILDI stands for targeted measurement of languages in the Internet from its French initials.

It is a new and ambitious project based on the creation of a program in capacity to extract language repartition and multilingualism parameters from any series of web sites, that with due consideration of the fact that a web site can have more than one language.

This consideration of web multilingualism in the process is both a complex problem and a historical breakthrough. It will allow, when applied to TRANCO, the series of the one million most visited web sites, to correct the extremely biased figures from W3Techs (see this article) and offer the first ever bias corrected measurement of language repartition in the webpages of the most visited sites.

The version 1 of MECILDI is the final testing process and soon we will share results, stay tuned!

This project first version has been funded by la délégation générale à la langue française et aux langues de France of France Ministry of Culture and will be next applied to obtain multillingualism characteristics and language percentages split for a series of gTLDs of languages of France (.alsace, .bzh, .corsica, .gp, .mq, .yt, .nc, .eus, .pf and .wf).

The post MECILDI appeared first on OBDILCI.