FREEDOM ON THE NET 2023 The Repressive Power of Artificial Intelligence Recommendations Policymakers, the tech industry, and civil society should work together to address the global decline in internet freedom. The following recommendations lay out strategies that policymakers, regulators, and private companies can adopt to prevent or mitigate illiberal uses of digital technology by both domestic and foreign actors, as well as the broader societal harms that the internet can exacerbate. While reversing the global decline in internet freedom will require the participation of a range of stakeholders, governments and companies should actively partner with civil society, which has always been at the forefront in raising awareness of key problems and identifying solutions to address them. 1. PROMOTE FREEDOM OF EXPRESSION AND ACCESS TO INFORMATION Freedom of expression online is increasingly under attack as governments continue to restrict connectivity and block social media platforms and websites that host political, social, and religious speech. Protecting freedom of expression will require strong legal and regulatory safeguards for digital communications and access to information. Governments Governments should maintain access to internet services, digital platforms, and circumvention technology, particularly during elections, protests, and periods of unrest or conflict. Imposing outright or arbitrary bans on social media and messaging platforms unduly restricts free expression and access to information. Governments should address any legitimate risks posed by social media and messaging platforms through existing democratic mechanisms, such as regulatory action, security audits, parliamentary scrutiny, and legislation passed in consultation with civil society. Other methods to address legitimate security problems include strengthening legal requirements for transparency, data privacy, and platform responsibility, such as mandatory human rights due diligence and risk assessments. Legal frameworks addressing online content should establish special obligations for companies tailored to their size and their services, incentivize platforms to improve their own standards, and require human rights due diligence and reporting. Such requirements should prioritize transparency across core products and practices, including content moderation, recommendation and algorithmic systems, collection and use of data, and political and targeted advertising. Laws should also provide opportunities for vetted researchers to access platform data—information that can provide insights for policy development and civil society’s analysis and advocacy efforts. Intermediaries should continue to benefit from safe-harbor protections for most of the user-generated and third-party content appearing on their platforms, so as not to encourage excessive restrictions that inhibit free expression. Laws should also protect “good Samaritan” rules allowing platforms to remove objectionable content in good faith, and reserve decisions on the legality of content for the judiciary. Independent, multistakeholder bodies and independent regulators with sufficient resources and expertise should be empowered to oversee the implementation of laws, conduct audits, and ensure compliance. Provisions within the EU’s Digital Services Act—notably its transparency provisions, data accessibility for researchers, a coregulatory form of enforcement, and algorithmic accountability—offer a promising model for content-related laws. 30 @freedomonthenet #FreedomOnTheNet