xAI for unsupervised ML in anomaly detection. Use case of LUCA Comms

Trusted AI is an emerging discipline that is gaining more relevance both in academic research and in industry as nowadays more companies are defining their policies and criterias for the usage of data and the development of support decision systems. This matter can be addressed, for instance, with the definition of what is known as AI principles that could serve as a guide for it, like it is been done in Telefónica.

Telefónica’s AI principles address several issues regarding those sustainable developments. They include the usage of fair models that do not discriminate against a protected group (such as discrimination by race) and the usage of techniques that contributes to an increased transparency of those models so the logic behind the decision taken by them could be explained to different audiences. It is also mentioned that all AI designs should be human-centric and that security and privacy are crucial components that should be considered by-design. Last but not least, all those principles should not only be applicable for the company internal developments but also for the usage of third party AI software.