Language is a central tool in the search for academic and professional information. This development in the availability of linguistic data has fostered a new discipline in multilingual data science where terminology and translation formalisms are essential in business and for academic scientific research. From our experience teaching in the field of philology, in the digital humanities (data science, Internet), the priority would be to start with a skills upgrade, but also by training in technical translation tools to identify, analyze and use translated texts, via queries on the web. We proposed to the 20 students of the University of Lyon (Bachelor level) to test the three online translation tools sequentially in order to compare the results. These translation tools (Google-Translator, Bing-Translator, and PROMT-Translator (https://www.online-translator.com)) are taught in the classroom for bibliographic research in English, following highlights:

• All students tested the three tools for a comparison on a natural language question: «Look for documents on the bird flu disease».
• 52% of them entered the question in natural language to have a correct translation.
• The ranking given by the students on these three translation engines is: (70% preferred Google, 25% for Bing and 10% for PROMT).

In conclusion, in the context of multilingual bibliographic research techniques, it is necessary to take into account the essential role of the specialty terminology that must be included in the modules of specialized training. The study shows that bilingual skills are needed for relevant access to scientific information.