Latest

AI Agents

Spain's Tax Agency is already training its officials to use AI

Spain's Tax Agency begins training its officials in AI with limited use and under institutional control

Spain's Tax Agency is already training its officials to use AI
Spain's Tax Agency is already training its officials to use AI

Spain's Tax Agency has begun preparing its staff to work with artificial intelligence (AI), in a process that is part of the digital transformation of public administrations. The topic was addressed at the nineteenth edition of the national conference of the Spanish Association of Tax Advisors (AEDAF), held in Madrid, where it was noted that the next step in the use of AI will be the development of agents capable of performing increasingly specialized functions.

During the meeting it was explained that the future of this technology in the tax field lies in the use of generative artificial intelligence to train agents with greater technical and operational capacity. The goal is for these systems to be able to support complex tasks within the tax administration.

Limited use of AI

For now, the use of artificial intelligence within the Tax Agency is limited. Internally, some officials already use AI-based tools, but with restricted functions. According to what was stated, these limitations are intended to prevent generative AI from being used in tasks that could jeopardize the operation of the administration or affect the rights of citizens.

Currently, AI is used mainly for support tasks such as assistance in drafting texts, translations, analysis of images or documents, preparation of summaries and transcription of information. To advance this technological implementation, the administration has put into operation four servers equipped with eight Nvidia graphics cards, purchased for 1.2 million euros.

What AI can officials use?

Given this scenario, a relevant question arises: can officials of tax administrations freely use any generative artificial intelligence tool available on the market, or must they limit themselves to those approved by their institutions?

The recommendation is clear: officials should use only the tools provided or authorized by the tax administrations themselves. This is due to the risks that can arise from the indiscriminate use of external tools, such as the possible violation of tax secrecy, the exposure of taxpayers' personal data or problems related to information security.

These aspects are linked to fundamental principles such as confidentiality, integrity and availability of data and systems. In addition, AI tools can reflect or amplify biases present in the data with which they were trained, reproducing stereotypes or generating biased responses, which represents another potential risk if they are not used under a regulated framework.

Importance of the recommendations

In this context, it is essential that tax administrations issue clear recommendations on the use of these technologies by their officials. Among them, it should be explicitly warned that AI tools that have not been previously approved by the tax administrations cannot be used. Likewise, staff should be instructed not to share information classified as confidential reserved or secret reserved through these platforms.

Respect for ethics and fundamental principles

The artificial intelligence tools used by tax administrations should comply with essential principles that guarantee lawful and ethical AI for the benefit of citizens.

Alfredo Collosa publishes a book on AI in the public sector

Among these principles, explainability, transparency, accountability and human oversight stand out. The protection of personal data and the duty of confidentiality must also be strictly respected.

In this process, the continuous training and education of public staff in the use of these tools will be key, also fostering interdisciplinary work and ongoing updating.

Not losing sight of the fact that AI is a tool

Artificial intelligence poses a scenario of disruptive changes for tax administrations in the medium and long term. Digital transformation does not only imply structural modifications, but also a cultural change within institutions.

Administrations must understand how AI impacts their functions and develop the necessary skills to use it efficiently. However, it should not be forgotten that AI is a tool to improve results, not an end in itself. Its implementation should not respond to a technological fad, but to a clear strategy regarding the benefits expected to be obtained.

In short, artificial intelligence should not be aimed at replacing public competences, but at expanding or complementing human capabilities, allowing officials to contribute greater value to their tasks and improve the quality and efficiency of public functions.

Alfredo Collosa

Author

Alfredo Collosa

Alfredo Collosa is a tax specialist. Public accountant. Bachelor's degree in Business Administration. Official Master's in Public Finance and Tax Administration from UNED/IEF in Spain. AMLCA Anti-Money Laundering Certificate (Florida University, US) and member of MetaLaw

The latest
From Huang to Botín, from Zara to Byron: corporate responsibility in global capitalism

Keep reading