The Government is going to collaborate with IBM to develop Artificial Intelligence language models in Spanish and co-official languages

Moncloa Palace, Madrid

The head of the Executive, Pedro Sánchez, accompanied by the Minister for Digital Transformation and Public Function, José Luis Escrivá, met with the senior vice-president of IBM and director of IBM Research, Darío Gil, where they discussed the new agreement reached between the Executive and the technology company to develop foundational Artificial Intelligence models based on Spanish and the co-official languages. The meeting was also attended by the Director of the Office for Economic Affairs and G-20, Manuel de la Rocha, and the President of IBM Spain, Horacio Morell.

This Memorandum of Understanding (MoU) with IBM - signed by the minister and the company's vice-president - launches the project announced by Pedro Sánchez at the last GSMA Mobile World Congress (MWC) Barcelona 2024 to build AI language models, specifically trained in Spanish and co-official languages, in open-source and transparent code, and with the intention of incorporating Ibero-American countries.

Pedro Sánchez stressed that the agreement will help to promote and position Spain as a leading country in developing ethical and responsible Artificial Intelligence and its roll out in the public administration, with the consequent positive impact on the different sectors of the economy and local innovation, particularly benefiting SMEs throughout the country. The Government will work to ensure that the public sector becomes a catalyst for innovation in the country.

The agreement also includes the promotion of research into techniques to accelerate the discovery and development of new materials that enhance the sustainability of the semiconductor industry and its value chain.

Language models and AI in administrations and SMEs

Specifically, the memorandum constitutes a collaboration agreement based on five strategic axes, which include the development of foundational models based on Spanish and co-official languages - both large and small-scale - that will in turn promote open strategic autonomy on AI and compliance with the Artificial Intelligence Regulation.

Foundational models in AI are systems trained with a large volume of data and information and that perform multiple tasks with high adaptability which, for example, facilitate the generation of content, reports and code or the automation of tasks, natural language processing, image classification, the creation of specific AI solutions, intelligent assistants or conversational systems, among others.

The agreement completes the development of a supercomputing platform for generative AI to accelerate the development of such models, including software based on open source frameworks, tools and infrastructure to enable hardware diversity. In this respect, the president of the Government of Spain inaugurated the MareNostrum 5 supercomputer at the end of last year; the first quantum computer in southern Europe that hosts the Barcelona Supercomputing Center - Centro Nacional de Supercomputación (BSC-CNS), which is already playing a fundamental role in promoting the reindustrialisation and strategic autonomy of Spain and Europe.

Furthermore, through the use of the foundational models and the generative AI platform developed, the agreement includes collaboration to bring and facilitate the adoption of Artificial Intelligence to public administrations and SMEs. It will also develop the ecosystem of AI creators and developers in Spain to position Spain as a leader in this field worldwide.

Spain, a world leader in ethical and responsible Artificial Intelligence

All strategic axes will be carried out under the framework of an open-source based Artificial Intelligence and with transparency and a responsible governance model as fundamental pillars, and a strategy based on the principles of open innovation and open science to offer greater economic opportunities and achieve higher levels of innovation.

Spain is an international reference in Artificial Intelligence after having promoted - during the Spanish presidency of the Council of the European Union - a pioneering regulation in this field and innovative initiatives such as the creation of the first Artificial Intelligence Oversight Agency at a European level. It has also been a pioneer in defending a global governance of digitalisation that protects citizens' rights with the Digital Rights Charter, an international benchmark and promoted by Spain since 2021.

Non official translation