Artificial Intelligence and the Environment: Ethical Challenges and Strategic Opportunities

Vocabulary flashcards covering key concepts, environmental metrics, technological frameworks, and regulatory initiatives discussed in the research on artificial intelligence and environmental sustainability.

Net-Zero Emissions

A state in which the total volume of greenhouse gas emissions added to the atmosphere does not exceed the volume removed or sequestered by the same organization or entity.

Carbon Dioxide Equivalent ($\text{CO}_2\text{e}$)

A standardized unit of measurement used to quantify carbon footprints by converting the global warming impact of various greenhouse gases into an equivalent quantity of carbon dioxide ($\text{CO}_2$).

Greenwashing

The deceptive practice of communicating vague, misleading, or unfounded environmental claims—or relying on external carbon offsets—to present an organization as ecologically sustainable without altering polluting operational practices.

Carbon Offsetting

A mechanism by which organizations or individuals compensate for their greenhouse gas emissions by purchasing certified credits that finance external projects, such as reforestation or renewable energy installations, that remove an equivalent amount of atmospheric emissions.

Water Security

The capacity of a population to safeguard sustainable access to adequate quantities of acceptable quality water to sustain livelihoods, human well-being, socio-economic development, and ecosystem preservation.

Online Nudging

Algorithmic interface choices deployed by e-commerce platforms to subtly persuade consumers into unintended purchases, resulting in indirect environmental costs such as excess packaging and frequent vehicle trips by gig-economy delivery drivers.

The Polluter Elite

A term referring to the wealthiest individuals and high-income nations whose consumption patterns produce disproportionately massive carbon footprints relative to the bottom $99\,\%$ of the world population.

AI Cannibalism

The phenomenon and operational risk of artificial intelligence models being trained on recursive data generated by preceding AI systems rather than fresh human-created data, leading to degradations in model output.

Model Distillation

A machine learning technique aimed at producing smaller, computationally efficient, and less energy-demanding models from larger foundational models to lower computational expense and resource consumption.

Shallow Neural Networks

Neural network architectures possessing minimal layers that can achieve high performance on specific known tasks using small datasets—such as approximately $10{,}000$ data points—substantially decreasing energy demands.

Digital Twin

A virtual, sensor-fed digital counterpart of a real-world system or municipality used by algorithms to model, monitor, and optimize complex processes such as traffic flow and urban energy consumption.

Smart Farming

Also referred to as digital agriculture or e-agriculture, the deployment of sensor data and artificial intelligence algorithms (such as regression models) to optimize irrigation, crop cultivation, and harvest planning.

GraphCast

An AI weather forecasting model developed by Google DeepMind, trained on $40$ years of historical meteorological data to generate accurate, lightning-fast medium-range global forecasts.

Project Natick

A Microsoft research and development project testing the feasibility and energy efficiency of submerged underwater data centers that utilize seawater for passive server cooling.

Adverse Digital Incorporation

A concept from information systems and ICT4D literature characterizing the inequitable or exploitative integration of vulnerable populations and Global South economies into digital systems and technological value chains.

TECTONIC

A research project funded by the European Commission aimed at applying artificial intelligence to seismic data to forecast earthquake locations, depths, and intensities.

Artificial Intelligence Environmental Impacts Act

Legislation introduced in the United States Senate on February 1, 2024, by Senator Edward J. Markey to require federal standards for investigating, measuring, and reporting the environmental footprint of AI systems.

AI Strategizing

The deliberate organizational planning and implementation of artificial intelligence systems to achieve strategic advantages, including environmental sustainability objectives such as emission monitoring and climate resilience.