High-performance AI without the risks. 

Guide to AI Auditing

Addressed to companies, public organisms, and citizens, this guide to auditing AI offers a general and replicable methodology. 

Specialized in AI auditing and algorithmic equity and justice and with a proven track record of real experience in audits, introduces the Guide to AI Auditing to make the current technology more explainable, transparent, and controllable. For this, the Guide offers a general and replicable methodology as an “instruction manual” so that citizens, public institutions, and companies have more control over the algorithms that are already deciding on their lives. 


The Guide to AI Auditing has three main objectives:  
    1. Protect the fundamental rights linked to privacy and protect personal data; 
    2. Provide clarity to the laws that apply to algorithmic systems; 
    3. Offer a methodology to ensure that these technologies are designed, developed, and used not only following the Law but also in a socially fair and responsible way. 


Big data, especially the one based on Artificial Intelligence (AI) that uses personal information, has an enormous effect on our daily lives. On a personal scale but also on a social, economic, and legal level. The ignorance of a large part of society (and the opacity of many of these systems) makes the systems not always follow ethical criteria. In these cases, the aforementioned systems may not apply ethical technology, something that translates into biases, bad practices, or discrimination. Especially, when we talk about vulnerable people. In some cases, because some AI systems are regulated by the General Data Protection Regulation (GDPR) also discriminates. That is why it is so important to evaluate the impact of the correct protection of personal data through some kind of external control. 


The content of the Guide to AI Auditing 

The Guide to AI Auditing is aimed at the people responsible for the use of algorithms, and data processing algorithm audits. Although, this guide is a tool that also seeks to reach the general public, increasingly interested in the effects of the application of algorithms in their daily lives. Therefore, it gathers both the definitions of the basic terms and the previous considerations necessary for their understanding, but also a more specialized content focused on the guiding principles of algorithmic auditing, the phases and recommendations for improving the systems after performing the audit. 


Algorithmic audits are a necessary way to make decision-making systems more explainable, 

transparent, predictable and controllable by citizens, public institutions and companies. 


Read here the Guide to Algorithmic Auditing in English.

Accede aquí a la Guía de Auditoría Algorítmica.