AI Ethics

The UNCTAD AI principles are a set of guidelines for the responsible development and use of artificial intelligence (AI). They were developed by the United Nations Conference on Trade and Development (UNCTAD) in consultation with a wide range of stakeholders, including governments, businesses, academics, and civil society organizations.

Machine Learning (ML)

Supervised learning Involves training a model on labeled data, where the algorithm learns to make predictions based on input-output pairs; unsupervised learning involves models that are trained on unlabeled data, and the algorithm discovers patterns or structures within the data; In reinforcement learning, agents learn to make decisions by interacting with an environment and receiving rewards or penalties for their actions.

Deep Learning

Deep learning is a subset of ML that uses artificial neural networks (ANNs) with multiple layers (deep neural networks) to model complex patterns and representations in data. Algorithms interpret sensory data through a kind of machine perception, labeling, and clustering of raw input. Common architectures include convolutional neural networks (CNNs) for image processing & video recognition tasks and recurrent neural networks (RNNs) for sequential data like time series or natural language.

Core Values

The UNCTAD AI principles are based on the following four core values

Human-Centeredness

AI should be designed and used to benefit all people, regardless of their race, gender, age, or socioeconomic status.

Inclusiveness and Equity

AI should be accessible and affordable to everyone, and its benefits should be shared equitably.

Transparency and Accountability

AI systems should be transparent and accountable to the people they affect.

Robustness, Safety, and Fairness

AI systems should be robust, safe, and fair, and they should be designed to minimize risks of harm.

Principles

The five UNCTAD AI Principles are:

Inclusive Development

AI should be developed and used in a way that benefits everyone, not just a select few.

Human Rights

AI should be developed and used in a way that respects and protects human rights.

Accountability

There should be clear mechanisms for holding people and organizations accountable for the development and use of AI systems.

Sustainability

AI should be developed and used in a way that is sustainable and environmentally friendly.

Transparency

AI systems should be transparent so that people can understand how they work and why they make the decisions they do

How to develop and use AI in a responsible and ethical way?

The UNCTAD AI Principles are a valuable framework for thinking about how to develop and use AI in a responsible and ethical way. They can be used by governments, businesses, and other stakeholders to guide their decisions about how to develop, deploy, and use AI systems. The UNCTAD AI principles are articulated in the following ten points

AI should be for the benefit of all. AI should be developed and used in a way that promotes human well-being and sustainable development.
AI should be inclusive and equitable. AI should be accessible and affordable to everyone, and its benefits should be shared equitably.
AI should be transparent and accountable. AI systems should be transparent and accountable to the people they affect.
AI should be robust, safe, and fair. AI systems should be robust, safe, and fair, and they should be designed to minimize risks of harm.
AI should be developed and used in a responsible and ethical manner. AI should be developed and used in accordance with human rights, international law, and ethical principles.
States should play a leading role in ensuring the responsible development and use of AI. States should develop and implement policies and regulations that promote the responsible development and use of AI.
The private sector should play a responsible role in the development and use of AI. The private sector should develop and implement ethical guidelines and best practices for the development and use of AI.
International cooperation is essential to ensure the responsible development and use of AI. States and other stakeholders should cooperate internationally to develop and implement policies and standards for the responsible development and use of AI.
Public awareness and education are essential for the responsible development and use of AI. States and other stakeholders should promote public awareness and education about AI, including its potential benefits and risks.
Research and development should be promoted to ensure the responsible development and use of AI. States and other stakeholders should promote research and development on AI, including on the ethical, legal, and social implications of AI.

UNCTAD believes that these principles are essential for ensuring that AI is used to promote sustainable development and inclusive growth. The principles are designed to be flexible and adaptable, so that they can be applied to a wide range of AI technologies and applications. Here are some examples of how these principles can be applied in practice

Inclusive AI

AI systems used in healthcare should be designed to be accessible to people of all income levels and backgrounds.

Human-centered AI

AI systems used in education should be designed to help teachers and students, not replace them.

Trustworthy AI

AI systems used in financial services should be transparent about how they make decisions, and should be accountable to human oversight.

Resilient AI

AI systems used in critical infrastructure, such as transportation and energy, should be designed to be able to withstand cyber attacks and other disruptions.

Fair AI

AI systems used in criminal justice should be designed to be unbiased and to avoid racial profiling

Sustainable AI

AI systems used in manufacturing and agriculture should be designed to minimize their environmental impact