They are invisible to practically all mortals, but nevertheless they have more and more social, political and economic influence. Algorithms have been controlling what appears to us every time we search for something on the Internet or consult a social network for years now, and their influence is increasing in a context where artificial intelligence (AI) is growing. His is a power that is not always studied.
A report prepared by the Digital Future Society (DFS) – an initiative promoted by the Ministry for Digital Transformation and Public Service, Red.es and Mobile World Capital Barcelona that connects policy makers, civic organizations, academic experts and businessmen to explore how technologies can be governed – confirms the need to advance and broaden the perspective in the evaluation of algorithms. In addition, it proposes an alliance between different sectors to make it possible.
The study recalls that the “impact of AI and the use of algorithms is something tangible.” It exemplifies that “the data and models may contain gender, race and other biases that lead to discrimination against some groups.” Furthermore, he points out that “AI systems have a high environmental impact, both due to the hardware and servers required and due to the enormous consumption of resources.” “How can algorithms be evaluated to detect the potential problems they contain or derive from their use, as well as contribute to their mitigation?”: is the question that guided a work in which UNESCO experts have been interviewed, the OECD, Fiddler AI and the European Parliament, 64 academic articles have been analyzed and 60 documents prepared by public bodies, third sector organizations, think tanks and other entities have been examined.
One of the main conclusions of the study is that we must “bet on an algorithmic accountability ecosystem that brings together the efforts of the public sector, the private sector and third sector organizations.”
The study of algorithms is not something new. But the DFS highlights the need to generate this alliance by betting on “more cooperative work” than what has been done until now. “Users, both inside and outside organizations, must also play an important role in the evaluation processes, and especially those groups affected by algorithmic decisions, as well as those who suffer greater vulnerability,” states the report.
