As a general-purpose technology, artificial intelligence (AI) already plays a role in our daily lives and work. However, its potential contribution to solving society’s most pressing challenges has yet to be realised. Laura Tyrer considers how AI might help us achieve our development targets, the ways the development sector might benefit from adopting AI technology and the influence civil society should have on ensuring AI remains a force for good.
AI research has made steady gains over the past 70 years, slowly manifesting what was once just science fiction. However, recent breakthroughs such as machine learning and natural language processing, have combined with better access to data and more powerful processing capacity to catapult AI towards transformational benefits for nearly every sector.
The social benefits of AI adoption are already evident. Advanced driver assistance systems demonstrate the potential to improve vehicle safety, while AI processes are applied in the healthcare sector for more accurate diagnoses. AI techniques are helping financial services providers identify fraud and corruption, and advancing weather prediction in the face of unpredictable climate change.
35%
15%
10%
Despite the evident value of AI, the World Economic Forum (WEF) has noted the development sector’s slow response in considering how AI might contribute to tackling the world’s most pressing challenges. Among the many possible interventions, AI could generate support for evidence-based decisions that align development sector projects with areas of need, inform policy decisions, coordinate development projects across countries and enable independent verification of international climate finance flows.
Having already transformed decision-making in industry, the WEF anticipates that AI is set to do the same for the development sector. With the Sustainable Development Goals (SDGs) deadline just seven years away, the question is the extent to which AI might contribute to SDG delivery.
What is artificial intelligence?
AI refers to the simulation of human intelligence using machine – usually computer – processes. It might be applied to software, natural language processing, speech recognition and machine vision. AI programming employs the cognitive skills of learning, reasoning, self-correction and creativity to absorb and analyse data, identify patterns and correlation, and then uses these to predict future states with which to generate new material.
Learning
AI acquires data and creates rules, or algorithms, to turn this data into actionable information. Algorithms give computing devices step-by-step instructions for how to complete tasks.
Reasoning
AI is programmed to choose the appropriate algorithm to achieve the desired outcome.
Self-correction
Algorithms are continuously fine-tuned to give the most accurate results.
Creativity
Neural networks, rules-based systems, statistical methods and other AI techniques allow AI to generate new images, text, music and ideas.
Demystifying AI
| 1. AI does not replace human intelligence. Limited to its programming and lacking the ability to understand context, meaning or emotion, it is incapable of truly creative, abstract or intuitive thinking. |
| 2. AI is well-suited to predictable, rules-based tasks. It cannot solve any problem that a human can solve. Guided only by data, it remains unsuitable for tasks requiring empathy, interpersonal skills and compassion. |
| 3. AI is not objective. Since the technology is only as good as the data it receives, algorithms developed off a biased data set will create equally biased AI, with the potential to exacerbate social inequities. |
| 4. AI cannot accurately predict the future. It can make predictions based on large amounts of data, but cannot account for chance, unforeseen circumstances and human decision-making. |
| 5. The fear that AI is a malevolent threat to humanity is unfounded. Widespread AI rollout carries risks and ethical implications, but how it affects humanity depends on its development and deployment. |
| 6. AI cannot solve all of humanity’s challenges. It may help us to solve some of our challenges but will only be as effective as the data we give it and the ethics of the policies and regulations that accompany its use. AI, while invaluable, cannot replace the need for human ingenuity, creativity and compassion. |
| 7. AI will not make humans obsolete. Although AI is likely to change the nature of work and be a disruptor in the short term, it is more likely that AI will be used to complement human abilities and enhance human intelligence, presenting the opportunity to create better solutions to human problems. |
AI and the SDGs – help or hinderance?
A 2020 Nature Communications study, The role of artificial intelligence in achieving the Sustainable Development Goals, assessed the role of AI in achieving the 169 SDG targets. It determined that AI has the potential to contribute to accomplishing 134 (79%) SDG targets across all goals, primarily through technological improvements. However, it also noted that AI could inhibit 59 (35%) of these targets.
The study established that AI could benefit societal SDGs by supporting the provision of food, health, water and energy services, while underpinning low-carbon systems and the creation of circular economies and smart cities with efficient resource use.
AI’s potential negative impact on societal targets pertains primarily to its carbon footprint. It is estimated that information and communication technologies’ electricity needs could account for 20% of global demand by 2030, compared to about 1% in 2020. For example, cryptocurrency applications are sustained by significant electricity generation, with Bitcoin requiring more electricity to operate annually than the 54 million population of Argentina.
Further hinderances to societal development might include the potential for AI to be used to manipulate societies, damage social cohesion, democratic principles or human rights. AI can reinforce societal biases if it learns algorithms from biased databases, inadvertently discriminating against women, minorities or other marginalised groups.
“AI’s potential negative impact on societal targets pertains primarily to its carbon footprint. It is estimated that information and communication technologies’ electricity needs could account for 20% of global demand by 2030, compared to about 1% in 2020.”
The study found that technological advances in AI have a potentially positive impact on the economic SDG targets. Its potential negative impacts were identified as increasing inequalities, impacting negatively on decent work, economic growth, industry, innovation and infrastructure.
While AI may be a significant enabler of environmental SDG targets in its ability to support climate change modelling, low-carbon energy systems and ecosystem health, research found that these could be undermined not only by AI applications’ high energy demands, but also its potential to drive resource exploitation.
The study urged that the rapid developments in the sector be supported by regulatory insight and oversight for AI-based technologies, to ensure that AI contributes positively to achieving the SDGs rather than an unequal and unsustainable AI-fuelled future. Not doing so would have implications for transparency, safety and ethical standards.
1: No Poverty
2: Zero Hunger
3: Good health and well-being
4: Quality Education
5: Gender Equality
Clean water and sanitation
7: Affordable and clean energy
11: Sustainable cities and communites
16: Peace, justice, strong institutions
Integrating AI into development efforts
For-profit sectors have leapt at the advantages AI and other modern tools offer. While slower to respond, the development sector stands to benefit too, with AI offering technology to advance project rollouts as well as tools to transform philanthropic operations and practices. Trialogue’s 2023 primary research found that while 35% of South African companies are investing in AI for their core operations, only 15% of non-profits are doing the same.
Alliance Magazine’s Philanthropy and AI webinar panellists identified generative AI as a powerful tool to build conversational mechanisms for grant applications and reporting. AI might be used to advance equitable practices in grantmaking, rapidly assessing large volumes of requests for those with the strongest alignment of priorities and outcomes. It could also streamline financial due-diligence practices. Natural language processing can be used to identify common themes across a range of reports and make connections to other data, delivering new ways to convey impact stories and build connections to new grant partners.
AI is already embedded in the software and other technology NPOs use and will be used by staff trying to ease their workload with or without organisational oversight. Authors of The Smart Nonprofit: Staying Human-Centered in An Automated World, Beth Kanter and Allison Fine, commented during a Lightful webinar, that AI presents a unique opportunity to reinvent and rehumanise the workplace, particularly in the non-profit sector.
“The social benefits of AI adoption are already evident.”
Intelligent and thoughtful application of AI has the potential to reduce the chronic capacity shortages in non-profits as well as the high levels of stress and burnout in the workplace that the pandemic amplified. Administrative overload and workflow burdens might be relieved by allocating appropriate tasks to AI, which could reconcile budgets, smooth workflow and make it easier to find information and intelligence in real time.
Chatbots could be employed to respond to frequently asked questions, freeing up staff to focus on more human, value-driven activities such as building relationships, working in communities and problem-solving.
One particular area they identified as primed to benefit from AI is fundraising. Fine maintains that AI presents an opportunity to challenge the negative trend of reduced donor giving, which has dropped 20% in the past 23 years. Non-profits might use AI to customise and personalise their donor engagements, transforming fundraising from transactional to relational and freeing up staff for direct human interactions.
CASE STUDY
AI in action
An AI innovation is resolving drug and medical equipment supply shortages at health facilities in Rwanda. Viebeg Technologies, a venture capital-backed health tech company, is helping healthcare facilities to procure supplies in real time. The platform uses AI to manage supply-chain processes, from shipping to warehousing, distribution and inventory management. It connects healthcare providers with manufacturers, removing brokers and middlemen from the value chain, reducing the cost to customer by as much as 40%.
In Kenya, irrigation and farming technology solutions company SunCulture combines intelligent hardware, Internet of Things, big data and neural networks to help farmers practice precision agriculture in the face of increasingly unpredictable climate patterns and maximise their yields at a lower cost. The company’s AgOptimized app draws on soil and weather data from soil sensors in the ground, local and Meteorological Aerodrome Report (METAR) weather stations, and meteorological satellites.
The data is analysed against historical climate change models using machine learning to provide farmers with detailed forecasts for their plots, together with guidance on planting, irrigating, fertilising and pest control.
Launched by Gro Intelligence, with support from The Rockerfeller Foundation, the Food Security Tracker for Africa tool provides free data and insight on factors related to food security across the continent. Users can navigate an up-to-date interactive map of Africa to access country-specific information on major crop production, crop stocks and demand, import price, drought conditions and crop health.
Combining research analysis and the scaling power of machine learning models, the tool may offer an opportunity to mitigate global food price inflation by increasing access to data and reducing fear-based reactions, while identifying critical at-risk areas. Indicating areas of food insecurity, the tool can be used by policymakers and aid organisations to better allocate limited resources.
When the SDGs were adopted in 2015, UN Human Rights sought to make human rights guidance available to inform the national policies needed to implement them. This took the form of the Universal Human Rights Index (UHRI), the world’s largest and most comprehensive human rights database of regional human rights systems’ considerations (customs, values, cultures and practices), showing the links between this guidance and the SDGs.
The index is now powered by AI. The UHRI shows the concrete linkages between individual recommendations issued by UN human rights mechanisms and the 169 targets of the 2030 Agenda. The new algorithm makes it possible to explore how human rights guidance can boost implementation of the SDGs by country, theme or group of affected persons.
Influencing AI for good
As AI advances are implemented with minimal regulation or discretion, could the development sector have more to consider than merely how to employ AI to their advantage?
European AI & Society Fund Director Catherine Miller maintains that civil society can and should play a more active role in shaping regulations that govern AI. She proposes that philanthropy, with its history of purpose-driven organisations oriented towards social impact, should participate actively in the discussion of how the technology is deployed and regulated.
“AI requires human oversight to understand, test and fact-check it, ensuring that it supports development.”
Given civil society’s interests, it has a moral responsibility to influence the ethical development of AI and ensure that a wide variety of voices have input into the design and development of the technology. In this way, civil society can help shape regulation to ensure that AI better serves people in society.
Such action requires substantial funding, however. Lobbying organisations would have to include those that understand technology, but also parties representing communities most affected by the impacts of AI and that can lend their social justice expertise to the debate.
The future of AI in the development sector
Philanthropy, rooted in the intention to challenge injustice and bring equity to marginalised groups, relies on emotion, empathy and human values. While this precludes any possibility of AI replacing philanthropic efforts, non-profits cannot and should not exclude AI from their work.
“While it is not a silver bullet for achieving NPO goals, AI can improve performance and enhance the delivery of social good.”
Kanter and Fine recommend that NPOs adopting AI ethically and responsibly will have the future advantage of greater adaptability and agility. They advocate for a human-centred approach to AI integration that aims to support and enhance staff efforts rather than replace people and that supports rather than undermines company culture.
Human connection should remain available and oversee AI-human interaction. Staff might be upskilled to work in tandem with AI and to take on more cognitively challenging tasks.
Every organisation should maintain AI ethical and responsible-use policies that include parameters guiding the use of AI and transparency guidelines, while remaining vigilant over the potential for AI bias that disadvantages already marginalised groups.
AI requires human oversight to understand, test and fact-check it, ensuring that it supports development.
A global standard for AI ethics
Recognising the pervasive way in which AI is reshaping the way we work, interact and live, and the need for thoughtful policies and regulations to ensure its ethical and responsible use, the United Nations Educational, Scientific and Cultural Organization (UNESCO) drafted the world’s most comprehensive international framework to shape the development and use of AI technologies.
The Recommendation on the Ethics of Artificial Intelligence was adopted by 193 member states at UNESCO’s General Conference in November 2021.
The Recommendation establishes a set of values in line with the promotion and protection of human rights, human dignity and environmental sustainability. It advances essential principles such as transparency, accountability and the rule of law online.
It calls for an end to the prevailing self-regulatory model that has prioritised commercial and geopolitical objectives over human wellbeing, calling on governments to establish the necessary institutional and legal frameworks to govern AI technologies and ensure they contribute to the public good.
Sources and further reading:
- Recommendation on the Ethics of Artificial Intelligence, UNESCO, November 2021.
- The role of artificial intelligence in achieving the Sustainable Development Goals, Nature Communications 11, 233, 13 January 2020.
- Artificial Intelligence and the Sustainable Development Goals, United Nations, 10 May 2022.
- The state of AI in 2022 – and a half decade in review, Mckinsey & Company, 6 December 2022.
- Cryptocurrency’s Dirty Secret: Energy Consumption, Columbia Climate School, 4 May 2022.
- TechTarget Enterprise AI.
- Debunking the Myths: 20 Misconceptions About Artificial Intelligence – Separating Fact from Fiction in the Age of AI, Ai-Scribed Insights, February 2023.
- Applying artificial intelligence for social good, Mckinsey & Company, 28 November 2018.
- Alliance Magazine Webinar, Philanthropy and AI, 10 May 2023.
- Lightful Webinar, Artificial Intelligence: what are the opportunities for nonprofits?, 3 October 2023.
- 3 strategies to leverage AI in the development sector, World Economic Forum, 7 October 2022.
Source: The original version of this article was published in the Trialogue Business in Society Handbook 2023 (26th edition).
