← Home
Field guide

AI and social responsibility

Norms on paper meet power, labour, and neighbourhoods—social responsibility is not automatic.

Overview

Social responsibility for AI spans human rights, fairness, transparency, and accountability in design and deployment, as well as who benefits from—and bears costs of—energy-intensive digital infrastructure. UNESCO’s recommendation on the ethics of artificial intelligence (2021) and the OECD AI Principles (2019) are widely cited reference points for values-based governance. Linking AI expansion to energy transitions also raises just-transition questions about communities, workers, and equity (World Bank, 2024).

By the numbers

Evidence highlights

3 beats
  1. 01

    United Nations Educational, Scientific and Cultural Organization (2021) centres human dignity, human rights, fairness, transparency, and accountability in its global normative framework for AI ethics.

  2. 02

    Organisation for Economic Co-operation and Development (2019) AI Principles emphasise inclusive growth, human-centred values, and risk management as part of trustworthy AI.

  3. 03

    World Bank Group (2024) just-transition framing highlights distributional impacts when economies shift infrastructure and employment—relevant when data centres and grids expand quickly.

Chart

Explore the data

Share of the population using the Internet (percent of population). The series starts from Our World in Data’s number-of-internet-users baseline divided by population for each country-year (ISO country codes only; regional aggregates are excluded). Add or remove countries to compare access—relevant context for who can participate in data-intensive services and AI ecosystems.

Loading chart data…

Trade-offs

Where it helps—and where it hurts

Strengths

  • Shared international language helps regulators, educators, and companies align due diligence and stakeholder engagement.
  • Strong labour and community safeguards can mitigate harms in supply chains for data, content moderation, and construction of energy infrastructure.
  • Participatory planning and benefit-sharing can improve legitimacy for major siting decisions when paired with enforcement.

Limits & trade-offs

  • Principles are not self-enforcing; implementation gaps vary widely across jurisdictions and firms.
  • Concentration of compute, data, and capital can widen global inequities in access, capability, and voice.
  • Fast deployment can outpace governance capacity on bias, privacy, and environmental justice without sustained investment in institutions.

References

Sources for this page

These entries are starting points for verification. Prefer the original report or dataset when checking numbers and figures.

  1. UNESCO AI ethics recommendationUnited Nations Educational, Scientific and Cultural Organization. (2021). Recommendation on the ethics of artificial intelligence. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000381137
  2. OECD AI PrinciplesOrganisation for Economic Co-operation and Development. (2019). OECD AI Principles: Overview. OECD. https://oecd.ai/en/ai-principles
  3. World Bank just transitionWorld Bank Group. (2024). Just transition for climate action. World Bank. https://www.worldbank.org/en/topic/climatechange/brief/just-transition
  4. Our World in DataRitchie, H., & Rosado, P. (2020). Electricity mix. Our World in Data. https://ourworldindata.org/electricity-mix (underlying grapher datasets include Ember and Energy Institute series, cited per chart metadata).