• AI adoption is growing in Spain at the same rate as cyber-attacks on models.
• Only 5% of Spanish organizations are prepared to apply AI securely.
• SLMs are emerging as an efficient and safe alternative to large models.

The conference “Secure and Efficient AI: strategies for introducing Artificial Intelligence in the enterprise”, organized today by DigitalES and CEOE, has highlighted the crucial role that AI plays in business competitiveness, as well as the urgent need to strengthen security, governance and efficiency frameworks to ensure its responsible implementation in all types of organizations.

• • • • Image gallery
      • See full video of the event here

The meeting brought together technology managers, cybersecurity experts, institutional representatives and executives from large companies, in a context marked by the unprecedented growth of AI adoption in Spain and the significant increase in cyberattacks targeting artificial intelligence models.

The conference was inaugurated by Miguel Sánchez Galindo, CEO of DigitalES, and Fátima Báñez, president of the CEOE Foundation.

Sánchez Galindo stressed that “artificial intelligence is a lever for innovation, but without cybersecurity it becomes a strategic risk”. Only 5% of organizations in Spain are prepared to address AI securely, he highlighted(Accenture: The State of Cyber Resilience 2025).

Fátima Báñez highlighted the importance of artificial intelligence for the competitiveness of Spanish companies, stressing that its structural adoption must be a priority, especially for SMEs. She stated that AI will not only transform processes, but will require deep organizational adaptation at all levels. In this regard, he remarked that continuous training is essential for digitalization to reach the productive fabric in a capillary and effective way.

PRESENTATION OF THE REPORT “ARTIFICIAL INTELLIGENCE AND CYBERSECURITY: RECOMMENDATIONS FOR A SECURE IMPLEMENTATION IN ENTERPRISES”.

At the cybersecurity roundtable, moderated by Blanca Bardin (PwC), the report prepared by the DigitalES “Cybersecurity and AI” Working Group, led by NTT DATA, was presented. The document warns about the increase in threats linked to AI and proposes a comprehensive security framework for companies of all sizes. The debate counted with the participation of Elvira García (American Tower), José Luis Rojo de Luque (EY), Juan Carlos Pérez (Nokia), David Sandoval (NTT Data) and Iago Soto (Quobis).

The experts stressed that the expansion of AI assistants and agents, capable of interacting with corporate systems and executing actions, makes it necessary to reinforce security controls from the design stage. Given that AI cannot be “made aware”, the experts stressed the need to establish clear guardrails, train the models properly and continuously monitor their results to avoid biases, deviations or misuse.

The risks and frauds based on identity theft, deepfakes or phishing, with a real reputational impact for organizations, were also alerted. In this context, employee training was identified as the first line of defense.

SMALL LANGUAGE MODELS (SLM), PROTAGONISTS OF THE TABLE ON GENERATIVE IA

The second panel of the day, dedicated to Small Language Models (SLM), presented the new DigitalES report -an annex to its Generative AI White Paper-, in a panel moderated by Jorge Martos (Indra Group), with the interventions of Borja Lizarraga (Accenture), Mafalda Vizcaíno (Deloitte Legal), Victoria Gómez (IBM) and Javier Hernández (Overlap).

The DigitalES report analyzes the emerging role of SLMs as an efficient and accessible alternative to large language models (LLMs).

The experts agreed that SLMs are a strategic way to democratize AI, enabling efficient, secure and scalable use in enterprises and public administrations.

In the discussion they agreed that it is not a matter of choosing between SLM and large models, but of combining them, using SLMs for more specialized, repetitive and closer to the user tasks, within a distributed AI approach.

It was also stressed that SLMs facilitate better data governance, greater traceability, control of results and protection of privacy, by working with more limited data sets. The adoption of this technology, they pointed out, should be approached as a business and regulatory compliance decision, prioritizing the training of people and appropriate change management, starting with simple projects that can generate value in the short term.

PUBLIC ADMINISTRATIONS AND STRATEGIC SECTORS: VISION AND CHALLENGES

During the conference, Miguel López-Valverde, Minister of Digitalization of the Community of Madrid, analyzed the role of the public sector and stressed that administrations must not only adapt to the new scenario posed by artificial intelligence, but also lead its ethical and responsible adoption, guaranteeing data protection, fairness in access to digital services and improved administrative efficiency.

The minister highlighted the role of artificial intelligence as a key tool to improve and simplify the provision of public services, both for citizens and public employees. He stressed the importance of responsible AI training, data securitization and support for SMEs to facilitate the adoption of these technologies, in a context in which the Community of Madrid is accelerating its digitalization initiatives, such as the Digital Shield to protect the region from attacks and technological threats.

In a final panel focused on AI, cybersecurity and competitiveness, speakers included Pilar Clavería (from the Spanish Banking Association, AEB), Yolanda Fernández (from the distribution association ANGED), Héctor René(INCIBE) and Enrique Maldonado (Hospital Gregorio Marañón), moderated by María Muñoz (Telefónica Tech). The speakers addressed the challenges of sectors such as banking and distribution, where AI is transforming critical processes, but requires robust security, data protection and governance frameworks.

Participants agreed that AI is transforming the relationship with the customer and the citizen, always as a means to improve service.

From the field of cybersecurity, the need for secure AI by design, with risk analysis, continuous monitoring and ethical values embedded in the models was highlighted, in a context where fraud and phishing remain the main threats. Training, talent availability and organizational change were identified as key challenges, as well as cross-industry collaboration and the role of DigitalES as a strategic partner to accompany the adoption of AI in enterprises.

GAINING CONFIDENCE IN THE IA, THE GREAT EUROPEAN CHALLENGE

The event was closed by Aleida Alcaide, Director General of Artificial Intelligence of the Government of Spain, who highlighted the value of the reports presented by DigitalES for their practical approach and their contribution to building confidence in the use of artificial intelligence by companies.

He stressed that the main challenge is to “gain that trust,” which requires involving both employees and management and clearly defining the objectives for using AI, in order to take advantage of its potential in areas such as productivity, healthcare and education.

Alcaide also warned of the risks associated with AI, especially in terms of bias, explainability and cybersecurity, and stressed the need to “accompany its deployment with governance and best practices”, in line with the guidelines recently published by the government to facilitate compliance with the European AI Regulation.

The director general recalled the importance of the National Green Algorithms Program, and highlighted the creation in Spain of a pioneering UNE standard for measuring the energy efficiency of algorithms. This standard makes it possible to evaluate and compare the environmental impact of AI solutions and move towards a more responsible use of these technologies, especially in the business and SME sectors.

• Download DigitalEs reports:
• Artificial Intelligence and Cybersecurity
• Small Language Models (SLM), Annex to the Generative AI White Paper