Carlos Martínez (Telefónica): “Only organizations with governed data will be able to generate impact with AI”

The appearance of generative artificial intelligence in 2022 with the help of OpenAI and ChatGPT revealed the need for a large amount of data to feed the models in its base, but although not just any type of data is valid, these must also be ‘governed’ to guarantee that they are suitable for use. “Only organizations with governed data can scale in the use of AI,” explained the director of Artificial Intelligence and Data at Telefónica Tech, Carlos Martínez.

In an interview with Europa Press, Martínez insists that if the data that feeds AI is not well organized, structured and of quality, it can result in the reproduction of biases and discrimination, offer incorrect information or become outdated, which in the end can “cause damage or cause problems to the organization that uses it.”

Data governance is the discipline that helps make data useful and generate valuable responses through the supervision of its quality and the definition of policies, processes and roles that facilitate adequate management and traceability of the data.

Thus, it allows “authorized people and models who need it to access the data, but in a safe and controlled manner, also guaranteeing compliance with current regulations,” says Martínez, who emphasizes its importance in a context marked by a greater implementation of AI agents in the business environment, capable of executing actions autonomously.

It is also an essential pillar for companies and public administrations build AI use cases that “transform organizational processes” making them more effective and efficient” and to generate an impact on the income statements, on the quality of products and services and on the levels of customer and user satisfaction.

For this reason, Martínez defends that “an organization’s data constitutes a unique competitive advantage”, and against the idea that data governance slows down innovation with artificial intelligence, he affirms that “it is empirically proven that only those organizations that have governed data can scale the use of AI and generate true impact in their income statements”.

Governance also plays a fundamental role in building sector data spacesenvironments that allow organizations in the same sector to share and use data securely and in accordance with regulations on the use and access to that data, that is, in a sovereign manner.

As the manager explains, the sectoral data spaces are part of the applied AI strategy of the European Union, which aims to build a single data market through the interconnection of different data spaces. “For Europe, they represent the opportunity to contribute to technological sovereignty and generate value in the economic sectors in which the continent has a leadership position,” said Martínez, in reference to health, industry, energy, mobility, agriculture or tourism.

“Spain is leading the construction of these spaces and from Telefónica, which we want to become the best access route to digital technologies, we are contributing to their development,” said the head of Telefónica, citing as recent examples projects such as the National Health Data Space of the Ministry of Health, the Agri-Food Industry Data Space DATAGIA promoted by ITACYL in Castilla y León, or other private data spaces such as the health data space of the Vithas Foundation.

Both governance and data spaces promote the development of sovereign artificial intelligence models, whose control is maintained by organizations, companies and governments, with the guarantee that critical data and infrastructure are governed by national jurisdiction in matters of security and privacy. That is, AI models that are fed and adjusted with native data, that are stored on local servers and processed in the territory itself, and that are subject to access, audit and regulatory compliance policies.