The Moncloa. 06/05/2024. What is the Artificial Intelligence Strategy 2024? [Prensa/Actualidad/Transformación Digital y Función Pública]

1. Strengthen the deployment of AI throughout the economy. The Executive is committed to strengthening capabilities to face a growing demand for artificial intelligence products and services in four areas: supercomputing, cloud infrastructure, AI language models and the need for talent.

2. Facilitate the application of AI in the public and private sector. The public sector must drive the adoption of AI to improve service delivery to citizens and serve as a catalyst for change in the private sector. With this objective, the General State Administration will promote, through an innovation laboratory, AI pilot projects and innovative solutions. Additionally, it will develop a common data governance model. To promote the development of AI in the private sector, the Government already has the Kit Consulting program in place and has expanded the Kit Digital program.

3. Promote transparent, responsible and humanistic AI. The objective is to achieve a broad degree of social consensus on the uses of artificial intelligence, its limits and the way of interacting between people and the developments made by AI. Spain, through the Spanish AI Supervision Agency (AESIA), created in August 2023, aspires to lead the use of responsible, safe and ethical artificial intelligence, and to establish a governance framework that ensures the highest levels of transparency and reliability.

What is a language model?

The RAE defines language models as computational models that try to represent the expressions or sequences of words of a language. In addition, they can obtain the probability of a sequence of words in a language and predict the words that follow those that precede them.

What does an AI language trained in Spanish mean?

A large language model requires a large corpus of linguistic data that allows learning the vocabulary and the relationship between the words that make it up. Generally, to build a language model we tend to use generic corpora of each language.

The percentage of resources in English reaches up to 90% in many current systems. Training in Spanish in the large existing AI models does not reach 5%, which is why the Government has opted for the creation of a language specifically trained in Spanish and co-official languages, with open and transparent code and with the intention of incorporating Ibero-American countries.

This open foundational model of generative AI developed in Spain is called ALIA, it promotes a family of models that will have more than 20% training in Spanish and co-official languages, and will handle at least four million words. Fifteen institutions specialized in the subject are participating in its development, the first results of which will be available before the end of the year. The President of the Government, Pedro Sánchez, announced the creation of this model in February 2024 during the GSMA Mobile World Congress (MVC) Barcelona.

The Executive has decided to have its own model to reduce biases and improve practical applications that companies and administrations can develop, such as intelligent assistants, conversational and dialogue systems or content generation models. The new initiative, a pioneer at European level, will provide an open, public and accessible language infrastructure. In addition, it will generate ethical and reliable standards that guarantee the protection of fundamental rights, intellectual property and personal data.

How will the security of AI be guaranteed in Spain?

The Government considers that cybersecurity is an essential element of trust for the technological transformation of the economy. For this reason, it is developing a cybersecurity law that provides a clear and complete framework to develop national cybersecurity by improving the protection of information systems, networks and data.

The Executive also proposes actions to promote innovation, collaboration and the adoption of artificial intelligence technologies in the field of cybersecurity through the National Cybersecurity Institute (INCIBE).

Furthermore, the Spanish AI Supervision Agency (AESIA), a pioneer in Europe, is key to defining good practices in the use of artificial intelligence and guaranteeing the evaluation of the models that are developed. It will also ensure compliance with harmonized AI standards set out in European law.

What is the application potential of AI?

AI is present in a multitude of applications and services, and citizens are already using it in their daily lives: internet search engines, personal assistants, household appliances, process robotization and automatic translators.

The Artificial Intelligence Strategy also describes its potential in different areas and areas of professional activity, including:

• Healthcare: AI modernizes healthcare by analyzing medical tests and creating patterns, as well as assisting in the detection of rare diseases and genetic disorders. Likewise, it contributes to the design of new drugs by reducing production times and costs.

• Teaching and learning sector: Allows the creation of multidisciplinary educational content, developing individualized learning, regardless of the student’s place of residence.

• Environmental: AI can have a positive impact in the fight against climate change, for example, by improving the safety, speed and efficiency of means of transport to reduce their polluting emissions. It also offers better management of energy networks and the climate efficiency of buildings.

• Administration and companies: AI can generate new products and services.

Are sustainable AI projects being developed in Spain?

Thanks to the Artificial Intelligence Strategy (ENIA) promoted in 2020, 18 artificial intelligence projects in the field of energy and 25 projects in the areas of the environment and rural areas have already been financed. They stand out among them:

• Projects to optimize water resources in irrigation and industrial consumption

• Improvement in the traceability of the agri-food chain

• The optimization of agricultural and forestry production

• The conservation of terrestrial and marine biodiversity

• The application of AI to the complete life cycle of the electricity sector

• Optimizing the location of solar panels and wind turbines for the production of renewable energy.