Inside this website

Everything after this paragraph is totally AI generated using ChatGPT free version first answer without modification. This is the sole and unique page using explicitly AI in this website.

What you will find on the OBDILCI website (AI generated)

This page is a synthesis of the content of the website produced by ChatGPT. It was written on the basis of the different pages and publications available on the website, then adapted and translated into OBDILCI’s three other working languages. It is intended as an entry point to facilitate the discovery of our work; for detailed results, methods and references, we invite you to consult the original pages.

Understanding linguistic diversity on the Internet

The Observatory of Linguistic and Cultural Diversity in the Internet (OBDILCI) is a non-profit organization dedicated to studying the presence of languages and cultures on the Internet. Since 1998, OBDILCI has been developing methods to measure this presence and monitor its evolution in a constantly changing digital environment.

The core of this activity is research: producing data, developing indicators, designing reproducible methods, and comparing results with existing scientific research. A significant part of OBDILCI’s research is published in peer-reviewed publications.

Measuring the presence of languages on the Internet

The main OBDILCI research program focuses on indicators of the presence of languages on the Internet. Since 2017, the Observatory has developed a model for estimating, for several hundred languages, their relative weight in Web content and relating it to the number of speakers and their access to the Internet.

This approach seeks in particular to answer an essential question: is the presence of a language on the Internet proportional to its demographic weight and to the proportion of its speakers who are connected to the Internet?

The database makes it possible to explore these indicators language by language and to compare situations. It also highlights phenomena such as the virtual over- or under-representation of certain languages and the relative productivity of their communities in terms of digital content.

Measuring Web multilingualism

The second major area of research is the study of multilingualism on websites. A website does not necessarily have to be associated with a single language: many websites offer several language versions. This reality makes measuring the presence of languages considerably more complex.

The MECILDI program is the computational response developed by OBDILCI to this problem. It makes it possible to analyze a given set of websites and identify the languages present, in order to measure in particular the percentage of multilingual websites, the average number of languages per multilingual website, and the overall rate of multilingualism. The program also makes it possible to study particular populations of websites, such as those corresponding to certain ccTLDs, gTLDs, or the most visited websites.

This approach makes it possible to go beyond methods that assign a single language to each website and may therefore introduce a significant bias into the assessment of linguistic diversity on the Web.

Studies devoted to languages and territories

The website also brings together studies devoted to particular languages or linguistic areas: the languages of France, French, Portuguese and Spanish, as well as research on relationships between languages in different areas of the Web.

These studies make it possible to move from global indicators to more focused questions: the presence of a language in different countries, the development of its Web versions, relationships between neighboring languages, the effects of e-commerce, territories, types of domains, or levels of economic development.

A recent series of reports thus explores Web multilingualism on the basis of large databases covering tens of millions of websites. It makes it possible in particular to study relationships between the main language of websites, multilingualism, economic activity and domain type.

The Internet, artificial intelligence and the future of languages

OBDILCI is also interested in the relationship between languages and artificial intelligence. The emergence of large language models and machine translation tools opens up new possibilities for languages, but also raises important questions concerning their representation in digital data, their ability to be processed by language technologies, and the risk of new forms of inequality between languages.

This reflection naturally extends the Observatory’s mission: measuring linguistic diversity in the digital environment no longer consists solely of observing Web content, but also of understanding how the technologies that produce, index, translate or analyze this content take different languages into account.

Publications, data and previous work

The website provides access to the scientific publications, reports, presentations, videos and blog articles produced by OBDILCI or related to its work. It also preserves the record of earlier projects that have contributed to the development of research on linguistic and cultural diversity on the Internet.

The collection thus makes it possible to follow nearly three decades of research, from the first observations on linguistic diversity on the Internet to current methods for large-scale measurement.

A website designed as an open resource

OBDILCI wishes its research to be available to researchers, institutions, language organizations, digital stakeholders, and anyone interested in the future of languages in the digital environment.

The website is therefore at once a presentation space, a documentary resource, a data exploration tool and a research laboratory. Its aim is not merely to answer the question “which languages are present on the Internet?”, but also to understand why their presence differs, how it evolves, and what these differences tell us about linguistic and cultural diversity in the digital world.

Exploring the different sections makes it possible to move progressively from this general overview to the data, methods, studies and scientific references that underpin it.