21 September 2020

UNESCO AI Ethics Report

Yet another set of principles for Artificial Intelligence ethics, this time from UNESCO. 

The report by UNESCO states 

1. Artificial intelligence (AI) is one of the central issues of the era of converging technologies with profound implications for humanity, cultures, societies and the environment. AI is already having impact across all sectors, and it is already transforming the future of education, natural and social sciences, culture and communication. These areas, along with the higher goal of promoting the respect of human rights and human dignity, along with a culture of peace, are core to UNESCO’s mandate. 

2. As with previous technological revolutions, AI has the potential to transform the future of humanity for the better and in favour of sustainable development. However, it can also bring downside risks and challenges, in particular derived from malicious utilization of the technology, that affects human rights, or from the fact that highly complex AI systems can widen substantially the already existing high inequalities and divides. In fact, these technologies, and the digital transformation have a “winner takes all dynamic” that needs to be addressed. The impact of the technology will depend on the way humanity frames it and masters it, and on the way it prioritizes the goal of leaving no one behind. This is where UNESCO’s role in promoting social inclusion and fighting inequalities, is of paramount importance at the global level. In order to sketch possible scenarios and unlock AI’s potential to grasp development opportunities, while managing risks, it is important to develop a more comprehensive understanding of how societies are transformed by disruptive technologies, such as AI. 

3. Digital transformation is happening at a moment where increased inequalities of income and opportunities are at its highest level in decades (within and between countries). These technologies are already contributing to deepening such inequalities in a profound way, as few firms and a handful of countries own the largest share of these technologies. This can change the landscape beyond recognition. Core elements in the AI life cycle determine who participates and who doesn’t. This can be related to data governance, which is a key issue (including data collection, data ownership, data sharing and data access). Data is the core ‘raw material’ that allows countries and firms to advance new solutions and products, but it is highly concentrated and there is no shared international governance framework. The skills necessary for data mining and the development of models and algorithms are also highly concentrated. Therefore, there is an essential link between data and modelling work, and the governance of AI. 

4. The inclusion of women in the digital transformation, or the lack of it, is especially pertinent, as the development of these technologies is massively contributing to opening an additional gender gap. Women do not participate equally in IT disciplines (they participate even less in IT than in science, technology, engineering and mathematics (STEM); they are underrepresented in the industry, and therefore in the development of AI systems. This opens a worrisome trend by which gender biases and stereotyping are being reproduced massively in the technologies. If technologies are not developed and applied in a gender-responsive way, they are therefore likely to reproduce and reinforce existing discriminatory social norms and even more violence against women (which is already happening). More women, including from different ethnic origins, must be involved in the design and deployment of AI tools. The same can be said about vulnerable populations, and those that are left behind, or discriminated against. They will find it more difficult to cope and benefit from the technological progress that AI technologies bring. 

5. This is why the work needs to be accompanied by an ethical reflection. It is clear that AI technologies are not value neutral, but inherently biased, among others, due to the data on which they are trained, and the choices made while training on the data. It is also influenced by the fact that AI machine decisions (particularly those based on Machine Learning) cannot be fully predictable or understood. Moreover, as AI is a distributed technology, whose current practical governance is spread across numerous institutions, organizations and companies, the reflection on its good governance requires a pluralistic, multidisciplinary, multicultural and multi-stakeholder approach, opening up questions about what type of future we want for humanity. This reflection needs to address the main challenges in the development of AI technologies related to the biases embedded in algorithms, the protection of people’s privacy and personal data, the risks of creating new forms of exclusion and inequalities, the issues of just distribution of benefits and risks, accountability, responsibility, impacts on employment and the future of work, human rights, security and risks of dual use. 


7. Because of its profound social implications, many organizations and governments are concerned about the ethical implications of AI. National, regional and other strategies and frameworks on AI are being developed. There is a growing number of reports and guidelines on AI and ethics – such as those by the Council of Europe, the European Union, the Institute of Electrical and Electronics Engineers (IEEE), the Organisation for Economic Co-operation and Development (OECD), the International Telecommunications Union, the World Health Organization (WHO), and the UN Secretary-General’s High Level Panel on Digital Cooperation – which have identified important principles for the design, development, and deployment of AI. UNESCO has been following these discussions closely, e.g. as a Co-Champion on AI of the UN Secretary-General’s High Level Panel on Digital Cooperation, an observer in the European Commission High-Level Expert Group on AI and in the Council of Europe Ad hoc Committee on Artificial Intelligence (CAHAI), a member of the OECD Expert Group on AI (AIGO), a partner of the ITU AI for Good Global Summit and a participant in other intergovernmental forums. UNESCO also hosted a high-level dialogue on the ethical dimensions of AI during the World Summit on the Information Society (WSIS) Forum in April 2019, and presented ongoing work on ethics of artificial intelligence at a high-level policy session during the virtual WSIS Forum in July 2020. 

8. UNESCO has a unique perspective to add to this debate given its strong comparative advantage thanks to its universality in membership and drawing on its multidisciplinary expertise. In this regard, the Organization can truly provide a global and pluralistic platform for dialogue on the ethics of AI, bringing together both developed and developing countries, different cultural perspectives, as well as various stakeholders within the public and private spheres. Therefore, in addition to the many ethical guidelines and frameworks that are currently being developed by governments, companies, and societal organizations, UNESCO can contribute to the development of AI for the benefit of all of humanity, sustainable development, and peace. For this purpose, UNESCO acts as a bridge between Member States – which have highlighted repeatedly their support for UNESCO’s AI work – and civil society, the technical community, academia, the private sector, including the cultural and creative industries, building on its record of multi-stakeholder consultation and consensus building. 

9. Building on the previous work of UNESCO’s World Commission on the Ethics of Scientific Knowledge and Technology (COMEST), in August 2018, the Commission was requested to prepare a preliminary study on the ethics of AI to help inform UNESCO’s reflection in this area. The work of COMEST on this topic provided the background for the preliminary study on the technical and legal aspects relating to the desirability of a standard-setting instrument on the ethics of artificial intelligence (AI), which was submitted to the Executive Board at its 206th session (document 206 EX/42), in accordance with Article 3 of the Rules of Procedure concerning recommendations to Member States and international conventions covered by the terms of Article IV, paragraph 4 of the Constitution. 10. Subsequently, the Executive Board adopted 206 EX/Decision 42 inscribing the preliminary study on the agenda of the 40th session of the General Conference. The General Conference examined the preliminary study and the comments and observations of the Executive Board thereon (document 40 C/67). By 40 C/Resolution 37, the General Conference recognized the benefits of an international standard-setting instrument on the ethics of artificial intelligence in the form of a recommendation (the “Recommendation”). Furthermore, Member States invited UNESCO to initiate the process of elaborating the Recommendation, and to submit to it at its 41st session the draft text of the Recommendation. The mandate provided by the Member States is highly important: there are many efforts trying to ensure governance of the development of AI, while this is the only standard-setting instrument that is clear on the fact that this governance should be based on strong respect of human rights, human dignity, and ethical considerations. 

11. The COVID-19 pandemic has exemplified how AI technologies can help, but also how much we need a global framework that informs the way we use them and the way we deploy them. To control the pandemic, we are giving access to our data and our activities, raising concerns regarding privacy and data protection and the temporal scope of extended powers, which all require reconceptualization to ensure respect for the rule of law. 

12. This international normative instrument will thus help the world reach an ethical consensus on AI, which is not just a technological turning point but also an anthropological disruption. This is a necessary step to more proactive thinking beyond the traditional legal approaches, which lag behind. The proposed Recommendation should become an ethical guiding compass and a normative bedrock allowing to build a strong respect for the rule of law in the digital world. THE PROCESS 

13. In line with the roadmap presented in Annex III of document 40 C/67, in March 2020, an Ad Hoc Expert Group (AHEG) (category VI meeting) was established to elaborate a first draft of the Recommendation. It is composed of 24 independent experts, appointed on a geographical balance basis from all UNESCO regions. From March to May 2020, due to the COVID-19 pandemic, the AHEG conducted its work using online/virtual modalities over a period of six weeks and produced a first version of the draft Recommendation. 

14. A multi-stakeholder consultation process on the first version of the draft Recommendation was conducted from June to August 2020. In light of the challenges posed by the pandemic, partners from around the world came together with UNESCO to support this consultation process based on three components: (i) public online consultation; (ii) regional and sub-regional virtual consultations co-organized with host countries/institutions in all of UNESCO’s regions; and (iii) open, multi-stakeholder, and citizen deliberation workshops organized by partners. The public online consultation received completed responses from more than 600 participants, generating more than 50,000 comments. Furthermore, eleven regional and sub-regional virtual consultations, including two with youth, were held in all of UNESCO’s regions, involving more than 500 participants. The majority of the consultations were opened at a high level, and included the Assistant Director-General for SHS. Finally, the open, multi-stakeholder, and citizen deliberation workshops drew approximately 500 participants. 

15. The consultations generated a rich conversation that raised awareness, triggered regional and sub-regional debates, and willingness of diverse stakeholders to collaborate with UNESCO in this endeavour. In at least three regions or sub-regions, participants have expressed the willingness to establish regional and sub-regional networks of different disciplines to continue discussions on the ethics of AI. 

16. The consultations have clearly shown the appetite to have an ethical framework on AI. Some regions have called for co-creation practices along the whole AI life cycle, ensuring broader understanding of education, ensuring gender equality, preserving the value of the native cultures beyond their languages, strengthen the need for human rights framework. They have also called for peace to be added as the only way to live in harmony and underlined the need for the document to be intergenerational in terms of values. There is evident strong acceptance of the environmental dimension of AI technologies. Moreover, there is support for its comprehensiveness in terms of including all the ecosystems and departing from a purely anthropocentric view. 

17. Consultation on the first version of the draft Recommendation was also held with the UN system through the intersessional meeting of the High-Level Committee on Programmes (HLCP) on 29 July 2020. All entities expressed their full support and recognized the importance of UNESCO’s work to develop this Recommendation and its relevance to the UN system, Member States and other actors. In general, it was emphasized that the text can and should be more ambitious by underlining more its innovative features and by proposing stronger language and far reaching recommendations when it comes to issues such as governance, human rights, and implementation after its adoption. This is necessary in order to effectively ensure ethical approaches to AI worldwide, to stand out from other documents in the field and to help position UNESCO as a reference for capacity building for Member States in the area of ethics of AI. The comments of UN partners were transmitted to the AHEG accordingly. 

18. The AHEG then revised the first version of the draft Recommendation from mid-August until early September 2020, based on the feedback of the consultations and on its further discussions. As such, the first draft of the Recommendation is transmitted to Member States together with this preliminary report. ... 

19. When preparing the first version of the draft Recommendation, the AHEG’s discussions concentrated primarily around the risks and challenges inherent to the rapid development of AI technologies that have been outlined at the outset of this report. Specific attention has been paid to how much AI technologies are already massively impacting our world, and to an urgent need for a common framework and narrative. The world is not fair and just, on the contrary, and we are often receding from our long-standing achievements. The COVID-19 pandemic is exacerbating the inequality trends by hitting hardest the most vulnerable population. The fundamental dichotomy is either the technologies contribute to worsen the problems of exclusion, inequalities and division, or they are used for noble purposes. Thus, the AHEG identified and conceptualized the issues that need to be corrected and the outcomes that they want to achieve with the proposed ethical framework. 

20. While recognizing the potential role of AI in contributing to solving complex challenges and bringing immense economic value, the AHEG stressed the concerns that AI creates regarding gender equality, diversity and inclusion, biases, privacy and data protection, increase of disinformation, digital exposure by populations at risk, just to name a few. The AHEG’s deliberations showed that today, at the global level, there is a need for a general universal ethical guidance in terms of core values that must underpin the development of AI systems. Due to their transnational character, durable solutions can be found only at the global level. Such an instrument should provide a global mechanism to document the social-cultural changes brought about by the rapid, non-linear development of AI and related ethical issues. 

21. Having agreed that AI technologies are not value neutral, the AHEG took the approach of identifying foundational values and closely interlinked principles, which inform further outlined policy actions. Foundational values have a role of necessary preconditions or prerequisites for principles and policy actions to work. Taken together they form a coherent framework to ensure ethical AI. The first version of the draft Recommendation has identified the values of human dignity, human rights and fundamental freedoms, leaving no one behind, living in harmony, trustworthiness, and protection of the environment. The latter has been purposefully elevated to the level of values because the analysis of various documents proposing principles for AI has shown that the protection of the environment receives little attention or is overlooked. 

22. The AHEG has also identified principles, which are associated with the human-AI systems interaction and those reflecting characteristics associated with the properties of AI systems themselves. The AHEG has recognized that there is an emerging convergence around the importance of certain principles that allow ensuring ethical approaches within the AI lifecycle. Therefore, the first version of the draft Recommendation includes such commonly recognized principles as human oversight, privacy, fairness, transparency and explainability, safety and security, responsibility and accountability, and others. Apart from that, the AHEG proposed new principles, which have not been previously seen elsewhere, such as proportionality, which helps ensuring that AI methods used correspond to the aims, do not bear excessive risks and are appropriate to the context. Nevertheless, the variety of source materials showed that both the notions and substance of principles may vary significantly, and thus the most challenging task for the AHEG was not identification of principles but clarifying their meaning and proposing policy actions that would ensure their implementation in practice. 

23. In view of implementing values and principles, the AHEG has proposed a number of concrete policy actions, which span from promoting diversity and inclusiveness to mitigating negative impacts on society, culture and the environment. The AHEG attempted to address the identified problems also by introducing innovative approaches, e.g. in the areas of governance, impact assessment and capacity building. To support that, the scope and addressees of the Recommendation have been expanded to include, among others, the private sector. 

24. The consultations have shown that the draft Recommendation can be bolder in terms of substance, proposed language and courses of action. This is especially so in terms of addressing gender issues, and broader inclusion, justice and integrity. Greater emphasis is needed to reflect the differential access to and impact of AI across subgroups and individuals, especially women, girls, older persons, youth and children, which should run through the entire document. It was suggested for the draft Recommendation to address more the issues of diversity of country groups as well as differences within countries in terms of level and diffusion of AI technologies. It was recognized that the draft Recommendation highlights the trade-offs, which are not always taken into account in other documents, and it was suggested to elaborate on those further in a more practical way. 

25. Suggestions were made to further elaborate on strategies to be followed in helping the low and middle income countries, including but not limited to countries in Africa, Latin America and the Caribbean, and Central Asia, as well as Small Island Developing States to access AI systems in various fields. There is a need for principles, actions and/or other recommendations aimed specifically at bridging gaps, capacities, resources, and power imbalances between countries and regions with regard to the proliferation of AI and capacity to regulate between AI-producing and AI-consuming countries. 

26. Suggestions were made in terms of the structure of the document. In particular, the policy actions have been perceived as rather complex and prolonged, which require further work. Similarly, aims and objectives of the document should become clearer and more concrete. This is important in order to keep the recommendation simple and pragmatic so that it can be operationalized quickly. Consultations also underscored that the additional value added of the UNESCO instrument is linked to the more action-oriented sections, advising on “how” to foster ethical AI developments, besides the “what” related to the definitions of what is ethical. 

27. Last but not least, it was suggested for the draft Recommendation to be more ambitious. This includes making bold proposals and be more assertive in suggesting that a stronger international legal framework is needed. It has been proposed to be more explicit about the innovative features that are reflected in this draft text in order for it to stand out among other documents. Additional innovative ideas and proposals were highly welcomed to appear in the final first draft of the Recommendation. 28. The comments received during the consultation process have been taken into account and have been reflected in the revised first draft of the Recommendation that is annexed to this preliminary report.

In noting 'Next Steps' the report states 

29. It is envisaged that the Recommendation, to be effective, would need to be accompanied by strong policy and capacity building support in its implementation and transformation into legal frameworks, and AI specific practices in the public and private sector. As such, UNESCO will be bringing its multidisciplinary competences to assist Member States in fulfilling the provisions outlined in the Recommendation once adopted. The Recommendation will also provide a solid foundation for the entire UN system in its collective response to the ethical challenges posed by AI technologies in various fields, with UNESCO playing a leading role in this area. In this regard, UNESCO has over 25 years of experience in developing and implementing normative work through policy advice and capacity building activities in bioethics and ethics of science and technology.

And of course 

UNESCO has the vocation to be the principal United Nations agency to promote and disseminate this Recommendation, and accordingly shall work in collaboration with other United Nations entities, including but not limited to the United Nations Secretary-General’s High-level Panel on Digital Cooperation, the World Commission on the Ethics of Scientific Knowledge and Technology (COMEST), the International Bioethics Committee (IBC), the Intergovernmental Bioethics Committee (IGBC), the International Telecommunication Union (ITU), the International Labour Organization (ILO), the World Intellectual Property Organization (WIPO), the United Nations Children's Fund (UNICEF), UN Women, the United Nations Industrial Development Organization (UNIDO), the World Trade Organization (WTO), and other relevant United Nations entities concerned with the ethics of AI. 

UNESCO shall also work in collaboration with other international and regional organizations, including but not limited to the African Union (AU), the Alianza del Pacífico, the Association of African Universities (AAU), the Association of Southeast Asian Nations (ASEAN), the Caribbean Community (CARICOM), the Caribbean Telecommunications Union, the Caribbean Public Services Association, the Common Market for Eastern and Southern Africa (COMESA), the Community of Latin American and Caribbean States (CELAC), the Council of Europe (CoE), the Economic Community of West African States (ECOWAS), the Eurasian Economic Union (EAEU), the European Union (EU), the International Association of Universities (IAU), the Organisation for Economic Co-operation and Development (OECD), the Organization for Security and Co-operation in Europe (OSCE), the South Asian Association for Regional Cooperation (SAARC), the Southern African Development Community (SADC), the Southern Common Market (MERCOSUR), as well as the Institute of Electrical and Electronic Engineers (IEEE), the International Organization for Standardization (ISO), and international financing institutions such as the World Bank, the InterAmerican Development Bank, and the African Development Bank. 

Even though, within UNESCO, the mandate to promote and protect falls within the authority of governments and intergovernmental bodies, civil society will be an important actor to advocate for the public sector’s interests and therefore UNESCO needs to ensure and promote its legitimacy.