Definition
In 2015, every member country of the United Nations agreed on 17 goals for a better world by 2030: no poverty, zero hunger, good health, quality education, clean water and more. Can AI help reach them?
The Sustainable Development Goals, or SDGs, are 17 global goals adopted by the United Nations in 2015 to end poverty, protect the planet and improve lives by 2030. AI for social good means using AI to make progress on problems like these, such as predicting floods, screening for disease, helping farmers and making education available in every language.
In simple words
The SDGs are like a to-do list for the whole world. AI can be one helpful tool for ticking items off, but only one. Many items also need money, good laws, roads, teachers and people working together.
Why it matters
AI's greatest value may lie not in apps for entertainment but in tackling hunger, disease, climate change and inequality. Knowing where AI genuinely helps, and where it does not, prepares you to use it for good, including in your capstone project.
Key terms
- Sustainable Development Goals (SDGs)
- The 17 global goals agreed by the United Nations in 2015, to be met by 2030.
- AI for social good
- Using AI to address social and environmental problems.
- Early warning system
- A system that predicts a danger, such as a flood, and alerts people in time to act.
- Screening
- Checking many people for signs of a disease so those at risk can get a full test.
- Precision agriculture
- Using data to give each field exactly the water, fertiliser or treatment it needs.
- Unintended consequence
- A harmful side effect that nobody planned.
Explained in detail
Examples across the goals
| Goal | How AI is being used |
|---|---|
| Zero hunger | Apps that identify crop pests and diseases from photos; forecasts of rainfall and yields |
| Good health | Screening chest X-rays for tuberculosis; detecting eye disease caused by diabetes from retina photos |
| Quality education | Translating lessons into Indian languages; practice apps that adapt to each learner |
| Clean water | Predicting where water pipes will leak; monitoring river pollution with sensors |
| Climate action | Flood forecasting and alerts; balancing electricity grids with more solar and wind power |
| Life on land | Identifying animals in camera-trap photos to track tigers and other wildlife |
Examples from India
Indian researchers and organisations have built AI tools for local problems. Teams have developed apps that help cotton farmers identify pests from photos of insect traps, so they spray only when needed. Hospitals and screening programmes have tested AI that reads chest X-rays to flag possible tuberculosis, helping in places with few radiologists. The government's Bhashini platform builds translation and speech tools for Indian languages so that digital services can reach people in their own language. Flood forecasting systems send alerts to phones in flood-prone areas days in advance.
Cautions
- Technology is rarely the main obstacle. A flood alert helps only if people have phones, trust the message and have somewhere safe to go.
- AI trained in one country may fail in another, with different crops, diseases, languages or skin tones.
- Projects for the poorest communities must be designed with them, not for them.
- AI uses energy and water, so its environmental cost must be weighed against its benefits.
- Data about vulnerable people needs strong privacy protection.
The most successful AI-for-good projects combine good technology with local knowledge, trusted local partners such as health workers or farmer groups, and a plan to keep the system running after the first excitement fades.
The goals are also connected. Better crop advice raises farmers' incomes (no poverty), improves nutrition (zero hunger and good health) and can reduce chemical use (life on land and clean water). A single well-designed project can help several goals at once. Equally, a careless project can help one goal while harming another, for example by using a great deal of energy, so good teams look at the whole picture. When you plan your capstone, start by naming the goal you want to help, then ask what success would look like for the people involved.
How it works
- Choose one SDG that matters in your area.
- Find a specific local problem under that goal.
- Ask whether data and prediction could genuinely help.
- List what else is needed: people, money, trust, infrastructure.
- Think about possible harms and who could be left out.
Try it yourself
A flood alert that worked
A river village in a flood-prone district receives AI-based flood forecasts.
Match the AI use to the goal
Which SDG does each example support?
- Spotting crop pests from photos
- Screening X-rays for tuberculosis
- Translating lessons into Indian languages
- Tracking tigers with camera-trap photos
Common mistakes
- Believing AI alone can solve big social problems.
- Using a tool built elsewhere without testing it locally.
- Designing for communities without involving them.
Check yourself
Summary
- The 17 SDGs are global goals for 2030, from ending hunger to climate action.
- AI helps with forecasting, screening, translation and monitoring.
- It works best with local people, trust and infrastructure, and must be tested locally.
Finished reading?
Sources: United Nations Sustainable Development Goals · ITU: AI for Good