We propose governing AI with the SDGs to ensure it advances global well-being. This alignment emphasizes ethical AI deployment, highlighting technology's potential to solve critical issues without compromising future generations. By integrating AI with the SDGs, we aim for a balanced approach that respects environmental, social, and economic pillars, ensuring AI contributes positively across all sectors.
Update on the occasion of the SDG Summit (18-19 September 2023)
The SDGs need rescue to get back on track. Unfortunately, little is expected to change as a result of the current discussions at the SDG Summit in New York (18-19 November). Once they encounter economic reality, the upcoming elections, and whatever drives politicians worldwide, noble promises and concerns regarding the SDGs will fall on the priority list. There is no political “constituency” for the SDGs. If you go outside the UN compound in New York, you’re not likely to meet many people who know about SDGs, let alone support them. It’s the truth.
What should we do? Not much. We shouldn’t give up, though. We need to be creative and think outside the box. For example, we should start guiding the future of AI with SDGs. The 17 SDGs and 167 targets make up the most up-to-date list of humanity’s values and top priorities. Additionally, they are also readily available and fully legitimate. Here, you can read about 15 reasons why the 17 SDGs should be used as guardrails for AI developments.
As AI continues to shape our reality, there are many calls to establish guardrails to guide future AI developments. Hundreds of AI ethical codes are outlined. New AI laws are proposed worldwide. Amidst all these initiatives, we are missing potential AI guardrails that are in plain sight: sustainable development goals (SDGs). The SDGs are the most comprehensive coverage of economic, human, and environmental development codified in 17 SDGs elaborated via 169 targets. They cover – among others – climate, health, gender, and inequality, from general principles to specific monitoring indicators.
So far, AI is mainly considered a technical way to realise the SDGs through the use of new applications and tools. Here’s why we should use SDGs as guardrails for AI governance.
Why are SDGs appropriate AI guardrails?
1. The SDGs are ready to be used. In AI governance, timing is crucial, as it was also stated clearly in the call for a 6-month pause in significant AI development signed by thousands of scientists and tech leaders. Other various calls for immediate action that describe AI as posing an ‘existential threat’ to humanity, highlighted this urgency for action. Any new development of AI rules and guardrails will involve lengthy and complex negotiation processes. Yet, SDGs can be deployed immediately as AI guardrails.
2. The SDGs have global legitimacy. At a time when it is challenging to have almost any global agreement, SDGs could be the only solution for AI governance with global buy-in. They have already been agreed upon by states in a legitimate setting. Through this legitimacy, SDGs can help build trust, which will be critical for future AI developments.
3. The SDGs are comprehensive in their policy coverage. They can address AI’s wide range of impacts on our core values, from respect for human lives and human dignity to the realisation of human potential via education and work. The SDGs are codifications of basic ethics and can replace hundreds of AI ethical codes that will take valuable time to create and agree upon.
The interactive illustration below shows how values at the heart of the SDG framework, such as affordability, inclusiveness, human-centeredness, and accountability (a total of 11 values), interconnect with AI.
4. The SDGs are operational and specific. SDGs can be operationalised in guiding AI developments through the matrix of goals, targets, and indicators. SDGs are an effective bridge between general principles and operational practices.
5. The SDGs are interdisciplinary. The main challenge of AI governance is connecting policy silos. The SDGs connect most policy fields, such as human rights, economic development, cultural values, and more. Metaphorically speaking, SDGs can provide guardrails not only for ‘one road’ – be it ethical or technological – but for the road network (see the interactive illustration below)
6.The SDGs are pro-development. There are major concerns that AI will trigger new and deeper digital divides for countries that do not have the means to participate in the AI race, which means most countries except the USA, China, EU, and a few other players. By applying the SDGs, we can address the risk of AI-driven divides within societies and globally.
7.The SDGs are inclusive. Most SDG targets for achieving inclusion could be applied to AI developments. Since inclusion is at the core of the SDGs in the societal development of youth, women, and other marginalised groups.
8. TheSDGs are diverse. Currently, AI is shaped by available data that comes mainly from developed countries and the Global North. In contrast, the SDGs were developed considering various cultural, religious, and local practices. The SDGs can guide AI development where it lacks the most – in diversity.
9.The SDGs are local. AI tends to have a very global and generic approach. Take ChatGPT that provides answers without taking into consideration local cultural contexts. Making AI more local is a major challenge, given the lack of local data and AI capacities in countries and communities worldwide. SDGs can help overcome this challenge like no other as they are anchored in the problems and concerns of citizens and local communities.
10.The SDGs are not centralised and imposing. Currently, a few companies are monopolising main AI developments and the AI market. The SDGs provide policy space for avoiding the centralisation and monopolisation of AI developments.
How can we deploy SGDs as AI guardrails?
11.AI itself can be used to develop AI guardrails. A rule-based AI model can be developed, starting with SDG targets and indicators. It can be enriched by text-based machine learning built around considerable texts and data gathered via SDGs reporting over the last eight years. Such an AI system could be developed transparently as an open-source project, enabling countries, citizens, and companies to monitor its development and deployment. Once in place, the AI model/system would be used as an ‘evaluation’ tool to determine whether specific AI solutions contribute to or go against particular SDGs.
Using SDG 5 (gender equality) as a guardrail, AI developers can reduce gender bias in AI platforms.
SDG 5: Goal and Targets
Indicators
5.1 End all forms of discrimination against all women and girls everywhere
5.1.1 Whether or not legal frameworks are in place to promote, enforce and monitor equality and non‑discrimination on the basis of sex
5.2 Eliminate all forms of violence against all women and girls in the public and private spheres, including trafficking and sexual and other types of exploitation
5.2.1 Proportion of ever-partnered women and girls aged 15 years and older subjected to physical, sexual or psychological violence by a current or former intimate partner in the previous 12 months, by form of violence and by age
5.2.2 Proportion of women and girls aged 15 years and older subjected to sexual violence by persons other than an intimate partner in the previous 12 months, by age and place of occurrence
5.3 Eliminate all harmful practices, such as child, early and forced marriage and female genital mutilation
5.3.1 Proportion of women aged 20–24 years who were married or in a union before age 15 and before age 18
5.3.2 Proportion of girls and women aged 15–49 years who have undergone female genital mutilation, by age
5.4 Recognize and value unpaid care and domestic work through the provision of public services, infrastructure and social protection policies and the promotion of shared responsibility within the household and the family as nationally appropriate
5.4.1 Proportion of time spent on unpaid domestic and care work, by sex, age and location
5.5 Ensure women’s full and effective participation and equal opportunities for leadership at all levels of decision-making in political, economic and public life
5.5.1 Proportion of seats held by women in (a) national parliaments and (b) local governments
5.5.2 Proportion of women in managerial positions
5.6 Ensure universal access to sexual and reproductive health and reproductive rights as agreed in accordance with the Programme of Action of the International Conference on Population and Development and the Beijing Platform for Action and the outcome documents of their review conferences
5.6.1 Proportion of women aged 15–49 years who make their own informed decisions regarding sexual relations, contraceptive use and reproductive health care
5.6.2 Number of countries with laws and regulations that guarantee full and equal access to women and men aged 15 years and older to sexual and reproductive health care, information and education
5.a Undertake reforms to give women equal rights to economic resources, as well as access to ownership and control over land and other forms of property, financial services, inheritance and natural resources, in accordance with national laws
5.a.1 (a) Proportion of total agricultural population with ownership or secure rights over agricultural land, by sex; and (b) share of women among owners or rights-bearers of agricultural land, by type of tenure
5.a.2 Proportion of countries where the legal framework (including customary law) guarantees women’s equal rights to land ownership and/or control
5.b Enhance the use of enabling technology, in particular information and communications technology, to promote the empowerment of women
5.b.1 Proportion of individuals who own a mobile telephone, by sex
5.c Adopt and strengthen sound policies and enforceable legislation for the promotion of gender equality and the empowerment of all women and girls at all levels
5.c.1 Proportion of countries with systems to track and make public allocations for gender equality and women’s empowerment
12.The SDGs should be complementary to other AI laws, standards, and declarations, including the OECD and UNESCO principles, the EU AI Act, Chinese AI regulations, and the Council of Europe’s planned convention on AI and human rights.
13.The environment, sustainability, and governance (ESG) principles are one channel to link the SDGs to the tech industry. Given the complementarity of SDGs and ESGs, the ESGs can be used to ‘translate’ SDGs to professional languages and cognitive framing used by businesses worldwide. By developing AI platforms, companies would also monitor their adherence to ESDGs.
14.The SDGs should be part of the evaluation of investments in AI, as this technology attracts a lot of capital. Investors should evaluate AI companies and projects based on their alignment with the SDGs.
15.The SDGs as AI guardrails should be communicated effectively. The prevalent view is that SDGs are some ‘do gooder’ approach with limited impact on ‘real life’. However, they are very realistic and practical governance tools. Thus, The UN, governments worldwide, and tech communities must put more effort into communicating the narrative of the SDGs as AI guardrails.
In essence, the SDGs are ready to be deployed to govern AI in timely, inclusive, informed, and impactful ways. It would be a win-win solution for all actors involved. By using SDGs as guardrails for AI developments, the tech industry can ensure its work does not go against core human values and principles underpinning SDGs and the ‘Agenda 2030’.
In addition, the SDGs’ concrete targets and indicators can be particularly useful in operationalising the general and abstract values and principles into the daily realities of AI developers, the tech industry, and policymakers.
At the same time, the SDGs would gain more prominence and relevance as they can become a valuable and practical tool for addressing one of the main challenges of our time: how to ensure that AI serves the core interests of humanity.
The next step in fostering this strategic interplay between SDGs and AI is the UN High-level meeting on SDGs to be held in September 2023 during the UN General Assembly.
