Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The impacts of artificial intelligence on environmental sustainability and human well-being

Noemi Luna Carmeno, Tiago Domingos, Daniel W. O'Neill

TL;DR

While the impacts of AI on income and health are expected to be positive, its impacts on inequality, social cohesion, and employment are expected to be negative.

Abstract

Artificial Intelligence (AI) is changing the world, but its impacts on the environment and human well-being remain uncertain. We conducted a systematic literature review of 1,291 studies selected from 6,655 records, identifying the main impacts of AI and how they are assessed. The evidence reveals an uneven landscape: 72% of environmental studies focus narrowly on energy use and CO2 emissions, while only 11% consider systemic effects. Well-being research is largely conceptual and overlooks subjective dimensions. Strikingly, 83% of environmental studies portray AI's impacts as positive, while well-being analyses show a near-even split overall (44% positive; 46% negative). However, this split masks differences across well-being dimensions. While the impacts of AI on income and health are expected to be positive, its impacts on inequality, social cohesion, and employment are expected to be negative. Based on our findings, we suggest several areas for future research. Environmental assessments should incorporate water, material, and biodiversity impacts, and apply a full life-cycle perspective, while well-being research should prioritise empirical analyses. Evaluating AI's overall impact requires accounting for computing-related, application-level, and systemic impacts, while integrating both environmental and social dimensions. Bridging these gaps is essential to understand the full scope of AI's impacts and to steer its development towards environmental sustainability and human flourishing.

The impacts of artificial intelligence on environmental sustainability and human well-being

TL;DR

While the impacts of AI on income and health are expected to be positive, its impacts on inequality, social cohesion, and employment are expected to be negative.

Abstract

Artificial Intelligence (AI) is changing the world, but its impacts on the environment and human well-being remain uncertain. We conducted a systematic literature review of 1,291 studies selected from 6,655 records, identifying the main impacts of AI and how they are assessed. The evidence reveals an uneven landscape: 72% of environmental studies focus narrowly on energy use and CO2 emissions, while only 11% consider systemic effects. Well-being research is largely conceptual and overlooks subjective dimensions. Strikingly, 83% of environmental studies portray AI's impacts as positive, while well-being analyses show a near-even split overall (44% positive; 46% negative). However, this split masks differences across well-being dimensions. While the impacts of AI on income and health are expected to be positive, its impacts on inequality, social cohesion, and employment are expected to be negative. Based on our findings, we suggest several areas for future research. Environmental assessments should incorporate water, material, and biodiversity impacts, and apply a full life-cycle perspective, while well-being research should prioritise empirical analyses. Evaluating AI's overall impact requires accounting for computing-related, application-level, and systemic impacts, while integrating both environmental and social dimensions. Bridging these gaps is essential to understand the full scope of AI's impacts and to steer its development towards environmental sustainability and human flourishing.
Paper Structure (22 sections, 6 figures, 2 tables)

This paper contains 22 sections, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Quantitative results
  4. Energy use and CO2 emissions dominate research on the environmental impacts of AI
  5. Subjective well-being is generally overlooked in social impact research on AI
  6. Systemic impacts are largely underexplored with respect to the environment
  7. Social impacts arising from the AI upstream supply chain are underexamined
  8. Environmental and well-being studies differ sharply in methods and scope
  9. Both environmental and well-being studies tend to rely on a narrow set of methods
  10. AI research suggests optimism about environmental impacts, but more concern for well-being
  11. Application‑level studies drive environmental optimism
  12. Sentiment is positive for some well-being impacts, but negative for others
  13. Qualitative results
  14. The explosive growth of LLMs and computing-related impacts drive environmental concerns
  15. AI may improve firms' environmental performance
  16. ...and 7 more sections

Figures (6)

  • Figure 1: Number of studies by impact type.(a) Environmental studies. (b) Well‑being studies. Energy use and CO$_2$ emissions dominate environmental studies, while the range of impacts considered by well‑being studies is more balanced.
  • Figure 2: Percentage of studies by impact category (computing-related, application-level, systemic).(a) Environmental studies. (b) Well‑being studies.
  • Figure 3: Percentage of studies by type of method and analytical scale.(a) Percentage of environmental and well-being studies by method used, classified as conceptual/theoretical, qualitative, or quantitative. (b) Percentage of environmental and well-being studies by scale of analysis, distinguishing micro-level studies (application or enterprise) from macro-level studies (national or global). (c) Percentage of environmental studies by impact category (computing-related, application-level, or systemic) and method used. (d) Percentage of well-being studies by impact type and method used.
  • Figure 4: Number of studies by specific methods and impact types.(a) Environmental studies. (b) Well-being studies. The vertical axis of both panels shows the specific method used, while the horizontal axis shows the impact type considered.
  • Figure 5: Percentage of studies by overall sentiment.(a) Percentage of environmental and well-being studies by overall sentiment on whether AI’s impacts are predominantly positive, neutral, or negative. (b) Number of environmental studies by overall sentiment through time. (c) Percentage of environmental studies by overall sentiment grouped by category (computing-related, application-level, or systemic). (d) Percentage of environmental studies by impact type and overall sentiment. (e) Percentage of well-being studies by overall sentiment grouped by category (computing-related, application-level, or systemic). (f) Percentage of well-being studies by impact type and overall sentiment.
  • ...and 1 more figures