TLDR: Microsoft is investing millions to significantly improve AI performance in European languages, addressing the current English dominance in AI models. The initiative includes establishing new research units in Strasbourg, France, by September, focusing on at least 10 EU languages, and openly sharing digitized linguistic data.
US tech giant Microsoft has announced a multi-million dollar investment aimed at significantly enhancing the performance of Artificial Intelligence (AI) models in European languages. This strategic move seeks to counter the prevailing dominance of English in AI development, ensuring that AI can better comprehend and serve Europe’s rich linguistic diversity. The company’s president, Brad Smith, emphasized that without such a course correction, ‘the survival of these languages and the health of these cultures is quite literally at stake.’
Starting this September, Microsoft will establish dedicated research units in Strasbourg, France. These units will focus on expanding the availability of multilingual data for AI development across at least 10 of the European Union’s 24 official languages, including Estonian and Greek. The work will involve extensive data collection, including the digitization of books and the recording of hundreds of hours of audio in these languages.
Smith clarified that this initiative is not about proprietary data acquisition. ‘This isn’t about creating data for Microsoft to own. It is about creating data for the public to be able to use,’ he stated, confirming that the collected information will be shared on an open-source basis with researchers worldwide. This commitment aims to make technology more inclusive for Europeans and aligns with the region’s growing push for technological independence.
The need for this investment stems from a critical imbalance: current leading AI models, primarily trained on English content, exhibit ‘less capability when it is in a language that has insufficient data.’ This deficiency can lead to accuracy differences of over 25 percentage points in certain European languages, such as Latvian, Greek, and Estonian, potentially compelling users to switch to English even when it is not their native tongue.
Microsoft’s efforts are also a response to the increasing concerns among European leaders regarding their dependency on US tech firms and infrastructure, especially following recent political shifts. The company has been actively positioning itself as a compatible partner in Europe’s drive for technological sovereignty, building on recent announcements concerning cybersecurity cooperation and data sovereignty measures for its European data centers.
Beyond linguistic AI, Microsoft is also contributing to the preservation of European cultural heritage. This includes projects like the digital replication of Paris’s Notre-Dame cathedral, which the company plans to gift to the French state, and the digitization of items from France’s BNF national library and Decorative Arts Museum. The Strasbourg initiative will also involve collaboration with the ICube laboratory at the University of Strasbourg, leveraging engineering capacity, Azure cloud credits, and the expertise of over 70 specialists from Microsoft’s international network.
Also Read:
- Europe’s AI Sector Sees Significant Investment Growth in H1 2025
- European AI Code Assistant Software Market Poised for Substantial Growth Through 2035
This comprehensive approach underscores Microsoft’s belief that AI systems must be tailored to serve the specific language, culture, and legal contexts in which they are used, moving beyond a neutral, English-centric model.
