Skip Navigation

Valencian Community, Spain
Create a Public Algorithmic Registry

Overview

Level of Government: Local

Lead Institution: Generalitat Valenciana

Challenge Area(s): Digital Governance


Description

				            				Reform Description

The General Directorate of Transparency and Participation of the Generalitat Valenciana is promoting the creation of GVA Confia-Registry of Algorithms of the Generalitat. "GVA Confia" is a computer application for the implementation of an algorithmic transparency policy. This tool, which takes the form of a catalogue, seeks to inform citizens about the use in the Valencian administration of algorithmic systems that have an impact on public decision-making or on the legal sphere of citizens. In other words, registration does not affect algorithmic systems based on their structure but on their purpose. "GVA Confia" seeks to offer information on algorithmic systems that include artificial intelligence, but also on those other algorithmic systems without AI, but that affect public decisions. 
In this way, the Registry of Algorithms of the Generalitat is not only a powerful tool for algorithmic transparency, but it is also a tool for accountability on the use of these technologies in administrative procedures or in public decision-making processes. 
In addition, this computer application is designed to comply with the general principle of compressibility and clarity, that is, providing the information as clearly as possible, ensuring that it is easily understandable using clear language. What's more, the algorithm registry is designed to provide information in the two official languages of the region. Another of the principles observed is that of maximum transparency, since it seeks to keep the restriction of access to information minimal and based on strict criteria of personal data protection or security. The principle of publicity has also been taken into account  , by virtue of which the administration must provide truthful information regarding its actions. On the other hand, the principles of citizenship orientation and continuity over time are transversal to the planning of this tool. Along the same lines, an attempt has been made to meet the principles of planning and evaluation, since "GVA Confia" was born as a pilot project that seeks to study the impact on society of algorithmic transparency and evolve over time, adapting to the needs of citizens and public administration. 


Problem(s) Addressed by Reform

Public Administrations are currently immersed in an unprecedented digitalisation process. In the 21st century, society demands a more agile, more effective and efficient Administration in the management and distribution of public resources. New computational capabilities, the progress of artificial intelligence applications and the increasing incorporation of professional profiles with digital skills in the Administrations are generating synergies that allow us to take advantage of the innovative potential to transform bureaucratic administrations into digital administrations. 
However, the implementation of new technologies, in the interests of efficiency, economy and administrative simplification, must also entail accountability for how this is done. Previous case studies have shown that computer systems are not without risk. Some examples such as the case of the Horizon system in the United Kingdom, the SyRI system in the Netherlands or BOSCO in Spain, have shown that the use of algorithmic systems can have a negative impact on fundamental rights, civil liberties or public decision-making. That is why algorithmic transparency stands as one of the fundamental pillars of transparency at this time. 
On the other hand, the lack of information on the use of new technologies, as well as the opacity of some of the algorithmic systems, generate distrust in the use of these applications for public decision-making since some sectors of the public are concerned about the lack of control or the massive processing of personal data . 
In addition, there is a regulatory vacuum, both national and international, on the ethical standards that algorithmic and artificial intelligence systems must comply with, especially in the public sphere. Moreover, the use of some of these systems conflicts with regulations on the protection of personal data and digital rights or on transparency.


With the development of "GVA ConfIA-Registry of Algorithms of the Generalitat", a system of transparency and accountability in the use of algorithmic and artificial intelligence systems is put in place. "Confia" means "trust" in Spanish. By offering detailed information in "GVA ConfIA" on the type of algorithmic systems that are used,  citizens' confidence in the use of this type of technology is reinforced, which results in a more efficient management of public resources. 
The information included in the Register of Algorithms of the Generalitat covers areas related to the architecture of the system, the code of the algorithm, the type of data it processes, the level of potential risk, remediation measures, information on licenses and patents, etc.  In this way, we achieve a more transparent and responsible public administration  with the implementation and development of these technologies. In addition, citizens' confidence in the institution and in the use of algorithmic systems in administrative procedures and public decision-making is strengthened. 
A more transparent, clearer, more accessible, more inclusive administration is a closer and more reliable administration. This facilitates accountability and promotes democratic values.


Relevance to OGP Values

The creation of the Algorithm Registry is relevant for transparency because it allows citizens to know what type of algorithmic and artificial intelligence systems the public administration uses for public decision-making and which of these have an impact on the individual or collective legal sphere. In addition, information is offered regarding what is the problem to be solved and how the algorithm contributes to the solution. Information on potential risks and contingency systems is also included. In short, it is not just a matter of offering a list of systems, but of offering all the information so that citizens know how they influence their lives.
On the other hand, GVA Confia has a direct impact on the accountability of the Administration of the Generalitat Valenciana. By providing information on algorithmic systems, an open citizen audit is being allowed. Both citizens and control bodies can check whether the algorithms are working correctly and whether they are doing so in a fair, equitable and inclusive way. In this way, the Administration of the Generalitat can also develop strategies for the responsible implementation of these technologies, especially those artificial intelligence systems that generate difficulties in the traceability of information.
Finally, in relation to civic participation, the Register of Algorithms of the Generalitat is a public transparency tool, with free and open access that will facilitate requests for access to public information on algorithmic systems. Offering a list of systems and explaining how they work allows citizens to be aware of how these systems impact public life, so it is easier to define what extra information is required in case you want to know in depth how the system works. In the event that systematic access requests are contemplated on the same topic, the inclusion of a field in the registry that offers this information on all algorithms is valued. 


Intended Results

By the end of the implementation period of GVA Confia-Registry of Algorithms of the Generalitat, we hope  to introduce all the algorithmic systems in the health field that influence medical decision-making. In addition, we hope that systems from different areas related to critical areas of care, such as the labor sector, will be registered in the Algorithms registry. 
We also hope that the algorithm registry will be expanded with more information related to the assessment of risks related to the protection of personal data or the identification of the competent bodies for its challenge. 


Milestones

The schedule proposed for this challenge is structured in X phases. Phase I is the design of an application in beta phase with the minimum fields required by transparency regulations (Law 1/2022, of 13 April, of the Generalitat, on Transparency and Good Governance of the Valencian Community). This prototype should be ready by March 2025. A group of professionals from the "Regional Office for the Application of Artificial Intelligence in Health of the Valencian Community" will test this application and their contributions will be taken into account for a first design of the application in a stable version. 
In Phase II , the first upload of algorithmic systems from the pilot project will take place in the application in its stable version. This application in a pilot project will be presented to society so that citizens know about the Algorithm Registry and start using it (May 2025). A first implementation evaluation report will be produced after the first six months (December 2025). 
Phase III will include more algorithmic systems that have an impact on public decisions (2026). In Phase IV , the possibility of expanding the functionalities of the Algorithm registry based on social and Administration needs (2027-2028) will be studied. 


Is Civil Society Involved?

Yes. 
Civil society is involved in shaping the action through the participation of the CIVIO foundation in the design of the pilot project. This civil society entity is made up of a multidisciplinary team that is responsible for analysing the transparency of Public Administrations in Spain and has carried out different research on public algorithmic transparency. CIVIO is the organisation that has taken the General Administration of the State of Spain to court to be able to access the source code of BOSCO, an algorithmic system developed by the Ministry of Ecological Transition for the granting of aid to vulnerable groups in energy matters and which has proven to have implementation flaws . 
We consider that the participation of this expert group is the key to developing a tool that is not only functional for the administration, but also for citizens and for the groups of experts in the field of algorithmic transparency in civil society. 
In addition, this collaboration has a vocation of continuity over time since it is expected that CIVIO will continue to participate in later stages of the development of the application.