|
||
|
TWN
Info Service on UN Sustainable Development (Sep26/05) UN:
AI dominance risks undermining right to development, warns study Penang, 11 Sep (Kanaga Raja) -- While artificial intelligence (AI) is increasingly shaping development priorities, labour markets, creative economies and knowledge production systems, without adequate safeguards, AI-driven economic concentration risks undermining participatory and equitable dimensions of the right to development by limiting who shapes technological systems and who benefits from them. This is one of the main conclusions highlighted by the Expert Mechanism on the Right to Development in a report (A/HRC/63/40) to the 63rd regular session of the UN Human Rights Council. The Expert Mechanism is comprised of five independent experts, providing the Human Rights Council with thematic expertise on the right to development in searching for, identifying and sharing best practices with Member States and promoting the implementation of the right to development worldwide. In its thematic study, titled “Artificial intelligence, cultural rights and the right to development”, the Expert Mechanism examines the implications of AI for cultural rights and the right to development. It recognizes the potential of AI to broaden access to creative tools, preserve cultural heritage and support linguistic revitalization, including for languages at risk of marginalization. However, it also warns that AI may undermine creativity, displace or devalue human creative labour, reproduce dominant cultural, aesthetic and linguistic patterns and weaken the conditions for diverse cultural expression. It said these risks are especially significant for developing countries, Indigenous Peoples, minorities and small- scale cultural producers, who often face unequal access to data, infrastructure, markets, regulatory protection and decision-making power. The Expert Mechanism recommends that AI governance be grounded in international human rights law and shaped by the right to development, such that AI supports inclusive, culturally diverse and human-centred development rather than cultural homogenization, digital dependency and the concentration of creative and technological power. According to the study by the Expert Mechanism, the right to development is indivisible and interdependent with all other human rights. It encompasses both individual and collective dimensions and requires equitable participation in decision-making processes affecting development trajectories. According to the study, Article 2 (3) of the Declaration on the Right to Development affirms that States have the duty to formulate appropriate national development policies aimed at the constant improvement of the well-being of the entire population and all individuals. “In the context of AI, this duty requires that innovation and economic efficiency do not exacerbate inequality, cultural homogenization or digital dependency.” AI & EQUITABLE DEVELOPMENT AI is increasingly positioned as essential to economic modernization and global leadership. This approach is reflected in most national AI strategies, according to the Expert Mechanism. However, it said the global AI ecosystem is characterized by asymmetries in data resources, computational infrastructure and capital investment. A small number of corporations and States exercise disproportionate influence over foundational models and digital platforms. Such concentration undermines international cooperation to eliminate obstacles to development, it added. The study pointed out that developing countries risk becoming providers of raw data while remaining dependent on proprietary technologies controlled elsewhere. It said this pattern may replicate historical forms of extractivism, this time in digital form, with cultural data and linguistic resources serving as inputs for AI systems without equitable participation or benefit-sharing. It said a right to development approach to AI governance therefore requires AI governance to address structural inequalities in access to data, infrastructure, skills, markets and regulatory capacity. It also requires international cooperation, technology transfer, capacity-building and participatory governance at both the domestic and the international levels. “In accordance with international human rights law, States bear the primary obligation to respect, protect and fulfil cultural rights and the right to development.” In the context of AI, it follows that the obligation to respect requires that States refrain from deploying AI systems in ways that directly interfere with cultural participation, artistic, academic and scientific freedoms, linguistic diversity or community control over cultural heritage and knowledge, said the study. The obligation to protect requires States to regulate private actors whose activities may adversely affect cultural rights, the Expert Mechanism stressed. The study said that given the central role of transnational technology corporations in developing and deploying AI systems, this duty entails establishing clear regulatory frameworks governing data collection, algorithmic design, intellectual property, competition, labour standards and access to remedy. “States must exercise due diligence to prevent, investigate, punish and redress abuses arising from corporate activities, including those with extraterritorial impacts.” It also said the obligation to fulfil requires proactive measures to facilitate equitable access to AI technologies and to support cultural ecosystems. According to the Expert Mechanism, such measures may include public investment in digital infrastructure, support for open and community-based AI initiatives, funding for cultural institutions and the promotion of multilingual AI systems that reflect linguistic diversity. The transnational nature of AI value chains raises complex questions concerning jurisdiction, accountability and access to remedy, it underlined. It said these challenges underscore the need for strengthened international cooperation, including harmonized standards, mutual legal assistance and inclusive global governance mechanisms. Recent years have witnessed a rapid proliferation of international and regional normative initiatives addressing AI governance, it noted. The study said at the global level, the United Nations Educational, Scientific and Cultural Organization (UNESCO) Recommendation on the Ethics of Artificial Intelligence adopted in 2021 constitutes the first comprehensive standard-setting instrument in this field. It articulates principles of human dignity, human rights, cultural diversity, environmental sustainability and international solidarity and affirms that AI systems should respect and promote cultural diversity, safeguard cultural heritage and support linguistic pluralism. The study said at the regional level, two instruments merit particular attention. The European Union Artificial Intelligence Act (Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024) adopts a risk-based, internal market-oriented approach, introducing obligations relating to transparency, accountability, human oversight and fundamental rights protection. It constitutes the first comprehensive binding legislative framework on AI within a regional legal order. The Council of Europe Framework Convention on Artificial Intelligence, Human Rights, Democracy and the Rule of Law of 2024 represents the first binding international treaty in this domain, requiring States Parties to ensure that AI systems are consistent with human rights, democratic values and the rule of law across their life cycle (preamble and arts. 4-7). These developments reflect an emerging consensus that AI governance must be grounded in international human rights law, according to the Expert Mechanism. However, it said that significant normative gaps persist, particularly with respect to equitable benefit-sharing, cross-border data governance and the protection of collective and community-based cultural rights. “Recent scholarship highlights structural imbalances in global AI development and calls for more inclusive and redistributive governance frameworks.” From a right to development perspective, technological governance cannot be confined to risk mitigation alone, said the study. Development entails equitable access to, participation in, and benefit from technological progress. In the context of AI, this means addressing structural inequalities in access to computational resources, data, infrastructure and markets. Without such measures, AI risks reinforcing existing global asymmetries and entrenching disparities between technologically advanced States and those with more limited capacities, the Expert Mechanism warned. The growing concentration of digital platforms and advanced AI capabilities in a limited number of corporate actors has significant implications for cultural diversity, it said. Notably, dominant platforms increasingly mediate access to cultural content through algorithmic curation and recommendation systems, shaping patterns of visibility and consumption. “These systems may systematically favour dominant languages, genres and commercially viable cultural narratives, while marginalizing minority cultures, less widely spoken languages and locally rooted forms of expression.” Market concentration in the digital economy also reshapes power relations within cultural and creative industries, said the study. Large platforms exercise significant control over distribution channels, data and monetization infrastructures, often resulting in asymmetrical bargaining relationships with creators. Independent artists and small-scale producers are often dependent on platform-based ecosystems that determine visibility and revenue streams, thereby constraining their capacity to participate on equitable terms. From the perspective of the right to development, such structural imbalances risk undermining fair access to the benefits of cultural production and participation in cultural life, according to the Expert Mechanism. Addressing these challenges requires a multi-dimensional regulatory response aimed at preserving pluralism in the AI-mediated cultural sphere, it suggested. “This may include the application of competition law to address excessive concentration, the introduction of interoperability and data access requirements, and the development of public interest-oriented data governance frameworks.” Support for decentralized, open or community-based digital infrastructures may contribute to a more diverse and inclusive cultural ecosystem, aligned with human rights principles, said the study. Highlighting the issue of data governance and cultural sovereignty, the study said AI systems also rely on the large-scale aggregation and processing of data, including cultural, academic, scientific, linguistic and artistic materials. Data governance frameworks, therefore, play a decisive role in determining whether AI-driven development is consistent with international human rights standards, particularly in relation to cultural rights, participation and non-discrimination. It said the conditions under which data are collected, accessed and utilized directly shape whose knowledge is represented, how cultural expressions are reproduced and who benefits from their use. For many Indigenous Peoples and marginalized communities, cultural data cannot be reduced to mere economic assets. Such data often embody collective identity, traditional knowledge, and spiritual significance, the study said. The unauthorized digitization, scraping or commercial exploitation of these materials, particularly in the training of AI systems, has increasingly been characterized as a form of cultural appropriation or “digital extractivism”, whereby value is derived from population knowledge without consent or compensation, it added. These concerns point to the importance of free, prior and informed consent, community-based data governance and equitable benefit-sharing. They also underline the need to distinguish between State regulatory sovereignty, Indigenous data sovereignty and community control over culturally significant knowledge. A human rights-based approach to AI governance therefore requires the recognition and operationalization of principles such as free, prior and informed consent, community-based data governance and equitable benefit- sharing, according to the Expert Mechanism. These principles are reflected in international standards concerning, in particular, Indigenous Peoples’ rights and collective self-determination, including the United Nations Declaration on the Rights of Indigenous Peoples, and are increasingly invoked in emerging frameworks on data sovereignty and Indigenous data governance, it added. The study said the transnational nature of data flows raises complex questions concerning regulatory sovereignty and development policy space. It said the global concentration of data infrastructure and AI capabilities risks enabling forms of asymmetrical extraction, in which culturally significant data originating in developing countries are processed and monetized elsewhere. “From a right to development perspective, this underscores the need for policy instruments capable of ensuring that domestic cultural, scientific and data resources contribute to local innovation ecosystems, capacity-building and equitable economic participation, rather than exclusively enriching foreign technology actors.” In summary, the study said the normative framework governing AI is evolving rapidly. Existing international human rights law already provides a robust foundation for evaluating AI systems through the lens of dignity, equality, participation and justice. The right to development offers a robust, human-centred framework for AI governance that integrates economic, social, cultural and political dimensions. It also requires not only the prevention of harm but also the proactive shaping of technological trajectories to advance inclusive, culturally diverse and sustainable development, it added. Regarding the implications of AI for cultural industries and the cultural labour markets, the study said that AI systems, particularly generative models capable of producing text, images, music and audiovisual content, are transforming the conditions of cultural production. “These technologies can lower barriers to entry, enable new forms of experimentation and expand access to creative tools. At the same time, they disrupt established professional roles, value chains and remuneration models within creative industries.” Generative AI systems trained on vast bodies of artistic and literary works can replicate stylistic features, generate derivative content and automate tasks previously performed by human creators. While such systems may positively assist human creativity, they may also reconfigure or displace elements of cultural labour, particularly in sectors characterized by precarious employment and freelance arrangements, the Expert Mechanism said. “The right to development requires that technological transformation be accompanied by measures ensuring equitable participation and fair distribution of the benefits.” In the context of creative industries, this includes ensuring decent work, fair remuneration and social protection for artists and cultural workers affected by automation, it added. Highlighting the issue of authors’ rights and generative AI, the Expert Mechanism noted that the large-scale ingestion of copyrighted works for the training of generative AI models has given rise to significant legal and ethical controversy. Developers frequently rely on data sets compiled through automated web scraping of books, music, visual art and audiovisual content, often without authorization or meaningful consent from rights holders, it said. The lack of transparency surrounding the composition of such data sets further complicates accountability and enforcement, it added. It said these practices raise complex and unsettled questions regarding the scope of permissible exceptions and limitations under copyright law, including text- and data-mining (i.e. the automated extraction of information from digital content) regimes, as well as the distinction between genuinely transformative uses and uses that may function as market substitutes. They also call into question the compatibility of large-scale data extraction with the obligation to respect and protect authors’ moral and material interests. A right to development framework does not preclude technological innovation; rather, it requires that such innovation be governed in a manner that ensures equitable participation in, and distribution of, its benefits, the study emphasized. “Relevant policy approaches may include collective licensing mechanisms; remuneration or compensation schemes financed by AI developers; enhanced transparency obligations concerning training data; and effective opt-out or opt-in regimes for creators,” it suggested. It said that given the transnational nature of digital markets and AI development, international cooperation may be necessary to prevent regulatory arbitrage (“forum shopping”), whereby actors exploit divergences between national legal frameworks. Creators in jurisdictions with weaker regulatory protections or enforcement capacities may face disproportionate exposure to uncompensated use and diminished bargaining power. Regarding labour rights and social protection, the Expert Mechanism said automation enabled by AI may affect a wide range of cultural occupations, including translators, illustrators, voice actors, composers, journalists and technical production staff. While new roles may emerge, such as AI trainers, prompt engineers and data curators, transitional disruptions could disproportionately impact vulnerable workers, it cautioned. “The right to development requires that structural economic change be accompanied by robust inclusive policy design, social protection and opportunities for re-skilling. Cultural workers should be meaningfully consulted and involved in decisions regarding the adoption and deployment of AI tools within their sectors.” While focusing on the risks posed by AI, the study said that AI technologies also offer opportunities consistent with cultural rights when appropriately governed. Examples include open-source tools enabling independent artists to access production capabilities previously limited to well-resourced studios; the AI-assisted restoration and digitization of deteriorating cultural heritage materials; and community-controlled language technologies supporting the documentation and revitalization of endangered languages. Such applications demonstrate that AI can serve cultural rights objectives when developed with community participation, transparency and equitable benefit-sharing, it said. “They also illustrate why a right to development approach should not be understood as anti-innovation, but as a framework for ensuring that innovation serves people, communities and diverse cultural ways of life.” In conclusion, the Expert Mechanism said AI presents both opportunities and risks for creative industries and cultural labour markets. It can democratize access to tools, foster cross-cultural exchange and stimulate new forms of artistic expression. However, without appropriate safeguards, it may undermine authors’ rights, exacerbate precarity and concentrate economic power, it added. The right to development provides a normative framework for navigating these tensions. It calls for participatory governance, equitable benefit-sharing and robust protection of moral and material interests, it further said. Overall,
the Expert Mechanism said AI governance remains at an early stage
of development. Current regulatory approaches, while evolving, are
fragmented and uneven in their capacity to address transnational and
structural impacts. The right to development provides a normative framework for AI governance, emphasizing participation, equity, accountability and international cooperation. It requires that AI governance address not only risk and innovation, but also the distribution of power, capacity and benefit, the study concluded.
|
||
