MODEL REGULATORY FRAMEWORK ON AI & MACHINE LEARNING IN SOUTHERN AFRICA PREAMBLE States should be committed to proportionate regulation of the risks of digital technologies. They can do so by exploring a range of mechanisms, including improving the policymaking process. These regulatory measures include norms, self-regulation, statutory codes of conduct, and rules in primary legislation. Given how digital technologies can quickly outpace law-making, States should also adopt non-regulatory tools to complement or provide alternatives to ‘traditional’ regulation, including industry-led technical standards. As digital technologies accumulate vast amount of data, they demand a distinct regulatory approach. They are a complex system whose creation and use relies on critical elements such as algorithms, access to data, and sufficient computing resources. To respond to these distinct challenges, States should take holistic, flexible, adaptable, transparent, and objective regulatory approaches that bring clarity by balancing the competing multi-stakeholders interests, including industry and civil society, as well as international commitments and obligations. This guidance focuses on the right to freedom of expression enshrined in Article 9 of the African Charter on Human and Peoples Rights, Article 19 of the Universal Declaration of Human Rights, and Article 19 of the International Covenant on Civil and Political Rights (ICCPR) and the right to privacy in Article 12 of the Universal Declaration of Human Rights and Article 17 of the ICCPR. 1. ETHICS AND LAW TO ENSURE THAT AI SERVES THE PUBLIC GOOD. States should firmly centre regulation on universal ethical principles and human rights as ultimate benchmarks for assessing the social acceptability of AI systems developed and deployed in and out of the African Union. 2. APPLICATION OF INTERNATIONAL HUMAN RIGHTS LAW. States have a three-fold obligation to respect, protect, and fulfil human rights and fundamental freedoms. 3. FREEDOM OF EXPRESSION The states shall take into account the preceding guidelines in adhering to the following principles in international law: a. Respect and protect everyone’s right to “seek, receive and impart information and ideas of all kinds, regardless of frontiers”. b. Whenever free speech is restricted through content moderation or otherwise, it must be provided by law and necessary for respecting the rights or reputation of others and protecting national security, public order, public health, or morals, for example, incitement and hate speech. c. In categorising hate speech and incitement speeches, intermediaries and states may find the following helpful categorisation: At the top is incitement to genocide, incitement to violence, and possibly to terrorism, and incitement to discrimination. This is followed by other forms of hate speech, including vilification, glorification, promotion, and justification, which are part of the pyramid below incitement. At the bot-