A/HRC/43/29 importance of fully considering relevant rules concerning the legality, legitimacy, necessity and proportionality of restrictions on human rights. The following sections highlight examples in the application of these key principles. A. Reinforcing equality and non-discrimination in new technologies Bridging the digital divide 42. The need to bridge the digital divide, which hampers access to technologies and its benefits, is recognized in the 2030 Agenda (General Assembly resolution 70/1, para. 15) and in several resolutions of the Human Rights Council. In its resolution 38/7, for example, the Human Rights Council called upon all States to bridge the digital divides, including the gender digital divide, and to enhance the use of information and communications technology, in order to promote the full enjoyment of human rights for all. 43. Assessing and addressing the digital divide requires attention not only to physical access to technologies and devices, but also to the different types of technology, the quality of access and the distributional equity of access. For example, while developing countries are gaining cheaper access to mobile technologies and the use of mobile phones has spread rapidly in most parts of the world, technology gaps are widening with regard to more advanced areas of technology, such as bandwidth availability. 24 The gender digital divide also persists, reflecting existing patterns of gender inequality and discrimination. Disaggregated data is needed to analyse and monitor the differentiated impacts of technologies in order to ensure equality and non-discrimination. Addressing bias in algorithms 44. While many dimensions of economic, social and cultural rights are to be realized progressively, States have an immediate obligation to ensure equality and nondiscrimination in law and practice. There is an urgent need to address the causes and impact of unintended bias and discrimination resulting from certain algorithmic and automated decision-making based on artificial intelligence and other technologies. Many algorithms tend to reinforce existing biases and prejudices, thereby exacerbating discrimination and social exclusion. Data-driven tools often encode human prejudice and biases, with a disproportionate impact on women and minority and vulnerable groups that are the subjects of those prejudices and biases.25 B. Legality, legitimacy, necessity and proportionality 45. Unless carefully regulated, the use of new technologies, in particular digital technologies, can easily lead to inappropriate restrictions on human rights. For example, big data and artificial intelligence, as well as digital identity systems, frequently depend on the collection and processing of data, often including massive amounts of personal data. This may amount to violations and abuses of the right to privacy when undertaken without the informed free consent of affected persons. Other rights often affected by the deployment of new technologies relating to economic, social and cultural rights include the right to freedom of opinion and expression, the rights to freedom of association and of peaceful assembly, and the right to an effective remedy. Restrictions on these and other rights must conform to the principles of legality, legitimacy, necessity and proportionality. 26 Limitations on a right, where permissible, must be necessary for reaching a legitimate aim 24 25 26 10 Human Development Report 2019 – Beyond Income, Beyond Averages, Beyond Today: Inequalities in Human Development in the 21st Century (United Nations publication, Sales No. E.20.III.B.1), p. 201. High-level Panel on Digital Cooperation, “The age of digital interdependence”, pp. 17–18; World Economic Forum, Global Future Council on Human Rights, “How to prevent discriminatory outcomes in machine learning”, white paper, March 2018. The Special Rapporteur on extreme poverty and human rights has pointed out that too many digital welfare state initiatives are characterized by a lack of attention to the importance of ensuring legality (A/74/493, para. 42).