AI: Transformative power and governance challenges

Can we ensure the inclusivity of artificial intelligence (AI)? What will AI mean for our economies or job markets? How can we make sure that AI doesn’t worsen current inequalities or introduce new ones? These questions call for a clearer understanding of AI and how it might affect our society.

“AI must benefit everyone, including the one-third of humanity who are still offline,” said UN Secretary-General António Guterres during “AI for Good” Global Summit in July 2023, which explored how AI can be used to achieve the Sustainable Development Goals (SDGs).

AI is present in everyday life, from booking flights and applying for loans to predicting pandemics or helping create inclusive environments for persons with disabilities. AI has the power to transform humanity if harnessed responsibly and made accessible to all.

Last week, during his lecture at the Royal Library of Belgium, UN Under-Secretary-General and United Nations University Rector, Professor Tshilidzi Marwala, spoke about the governance challenges of AI. As AI systems become more capable and autonomous, we must consider how to regulate their design, training, and use to align with human values and prevent unintended harm.

What do we have in mind when we talk about AI?

According to the World Intellectual Property Organization (WIPO), AI is generally considered to be a field within computer science aimed at developing machines and systems capable of performing tasks that typically require human intelligence, such as perception, language interaction or problem-solving. AI is founded on algorithms, which are translated into computer code containing instructions for the rapid analysis and transformation of data into conclusions, information, or other outputs.

One of the AI examples of a natural language processing tool is Chat GPT. It allows users to interact with the GPT-3 (Generative Pre-trained Transformer 3) model using natural language. The model is trained on a large amount of data, allowing it to generate human-like responses to various inputs. However, these technologies still have limitations, such as incorrect responses, high carbon footprint, lack of updated information and requiring access to the internet. Notwithstanding, they are gaining in popularity – launched in November 2022, Open AI’s ChatGPT-3 platform reached 100 million users by January 2023, making it the fastest-growing consumer app in history.

Positive impact of AI: SDGs accelerator

AI holds the potential to drive progress across all 17 SDGs, offering solutions to pressing global challenges. Its applications enable innovation, improve risk assessment and planning and allow faster knowledge sharing. On 26 October 2023, the UN Secretary-General launched the new UN Advisory Body, consisting of 39 experts from across the world, who will aim to harness AI for the common good. They will make recommendations by the end of the year on the areas of international governance of AI; shared understanding of risks and challenges; and key opportunities to leverage AI to accelerate the SDGs.

AI is already offering incredible possibilities. Below, we have included some of the key areas where AI has demonstrated significant progress and the capacity to bring about transformative change.

Healthcare:

AI has the potential to strengthen health service delivery to underserved populations; enhance public health surveillance; advance health research and the development of medicines; support health systems management and enable clinical professionals to improve patient care; and perform complex medical diagnoses. For example, AI is more accurate when it comes to designing personalised treatments or predicting heart attacks. When AI is combined with human diagnosis, the error rate is only 0.5 percent — compared to 3.5 percent for human doctors. According to the World Health Organization, non-communicable diseases cause approximately 89 percent of premature deaths in the European region, thus better diagnostics could help save millions of lives.

Doctor using a tablet — AI has the potential to strengthen health service delivery.

Environment:

AI can play a role in tackling environmental challenges, from designing more energy-efficient buildings to monitoring deforestation, optimising renewable energy deployment, or calculating the environmental footprints of products.

Humanitarian Aid:

AI can facilitate the analysis and interpretation of vast and complex humanitarian datasets to improve projections and decision-making. Mobile applications, chatbots and social media can create immediate feedback loops with people affected by humanitarian crises. Chatbots are being used to understand the needs of refugees in different languages. Digital cash can provide rapid and flexible assistance. Biometrics can help establish digital identity and reconnect families.

Education:

AI has the potential to address some of the biggest challenges in education today, as well as to innovate teaching and learning practices. AI can be used to help students learn new languages or to support children with disabilities by providing them with personalised resources.

Agriculture:

AI could transform agri-food systems and help end hunger. More efficient production technologies have the potential to provide stable incomes to rural populations while ensuring responsible use of environmental assets.

However, existing and emerging AI technologies are being rapidly deployed without a full understanding of how such AI systems may perform. As with any technology, AI holds promises… but also risks.

AI governance challenges

At the launch of the High-Level Advisory Body on Artificial Intelligence, the UN Secretary-General listed the potential harms of AI, expressing concerns over misinformation and disinformation; the entrenching of bias and discrimination; surveillance and invasion of privacy; fraud, and other violations of human rights. He said: “Without entering into a host of doomsday scenarios, it is already clear that the malicious use of AI could undermine trust in institutions, weaken social cohesion, and threaten democracy itself.”

While speaking at the Royal Library of Belgium, Professor Tshilidzi Marwala, who is also a Member of the United Nations Scientific Advisory Board and former Vice-Chancellor of the University of Johannesburg, shared insights on ethical frameworks, technical standards, and policy interventions needed to steer AI in a responsible direction. As potential challenges, he mentioned:

Computation (A process in which a calculating machine learns algorithms, which is expensive and has a high carbon footprint, for example, to train ChatGPT costs 100mln USD)
Ethics
Cross-border data flow (Data circulation during a pandemic or in relation to climate change should circulate freely, so we can be better prepared)
Access to data
Synthetic data (In order to be interpreted correctly, it is important that data is collected in a uniform manner)
Sensors (They help process signals and their resilience is important for our infrastructure)
Wars (AI can help to predict conflicts)
Truth (How to tackle the issue of software that learns to bluff?)
Human rights challenges (Technology should be used to advance human rights culture)

The Governance Challenges of Artificial Intelligence: A Lecture by Tshilidzi Marwala, UNU Rector and Under-Secretary-General of the United Nations; Royal Library of Belgium, 24 October 2023 Photo: UNU-CRIS

Ethical guardrails

Some of the main ethical issues of AI mentioned by Prof. Marwala included: privacy and surveillance; bias and discrimination; transparency and explainability; job displacement; security; accountability; regulation and oversight.

AI technology brings major benefits in many areas, but without ethical guardrails, it can threaten fundamental human rights and freedoms, for example image generators and video deepfakes present a substantial challenge to the preservation of information integrity. AI and automation can also have a greater impact on jobs in smaller cities and could deepen inequalities within societies by eliminating some functions.

Futureproofing

In a world where half of the population still cannot count on a stable internet connection: “We must work together for AI that bridges social, digital, and economic divides, not one that pushes us further apart,” said the UN Secretary-General in his remarks to the Security Council Debate on AI in July 2023. AI has the potential to advance the achievement of the Global Goals. However, we need to establish a consensus on regulation to guide the development and implementation of AI. This regulation should ensure that AI benefits all of humanity, fosters equity, and minimises disparities and biases.