AI & freedom of expression: OSCE-RFoM releases policy manual

by Eric Munch (Observer) posted on 31 January, 2022   (public)

SAIFE policy manual finds urgent need  to address surveillance-based advertising ecosystem to protect freedom of opinion and expression


On 26 January 2022, the Office of the OSCE Representative of Freedom of the Media (RFoM) released a policy manual, as part of their project “Spotlight on Artificial Intelligence and Freedom of Expression” (SAIFE).


The manual is the product of two years of research, the results of which are a set of human rights-centred recommendations with the aim of identifying human rights due diligence measure and procedural safeguards to address both individual and societal risks arising from an unwarranted use of AI in content governance. As its title indicates, the policy manual aims to provide technical, actionable and human rights-centric recommendations to safeguard freedom of expression in the context of AI use in content moderation and curation. To do so, it provides guidance to OSCE participating States on human rights safeguards with regard to developing and deploying AI for shaping and arbitrating online information spaces.




Among the numerous recommandations the manual makes to the OSCE participating  States, it defines the following elements as key for the ethical use of AI in content curation and moderation:

  • The necessity for algorithmic transparency
  • The need for interdisciplinary research, transparency of intermediary policies and content curation practices as preconditions to put media pluralism and diversity at the centre of algorithm design and deployment
  • The urgent need to address the surveillance-based advertising ecosystem to genuinely protect and promote freedom of opinion and expression in the digital age


The findings originated from expert workshops focusing on four primary areas of concern: security threats, hate speech, media pluralism and surveillance capitalism. The recommendations of the group on media pluralism might be of particular interest to EPRA members. Inter alia, they highlight that States should:

  • Provide support and resources to existing independent media NRAs that use a process of co-creation and inclusion of all national media actors and experts
  • Strengthen independent media NRAs (...) and involve them in public oversight and research
  • Eastablish sustainable media and digital literacy programmes for all societal groups


The workshops brought together 120 experts from various backgrounds to work on individual right to freedom of expression and opinion, and rights on a societal level including media freedom.


Source: OSCE Website

Further EPRA-related information:

EPRA's AI & Regulators roundtable, launched on 15 October 2021, enables EPRA members to discuss the use and development of AI-based tools by and for regulators, and to learn from each other. The group meets every 3 months to exchange updates and discuss ongoing projects of regulators or other stakeholders.