Anthropocentrism and Environmental Wellbeing in AI Ethics Standards

Vocabulary flashcards covering key definitions, philosophical ethical frameworks, and empirical findings from the scoping review on anthropocentrism, human-centred AI, and environmental ethics.

Anthropocentrism (Moral Philosophy)

The view that only humans hold intrinsic moral value or, at the very least, that humans possess more intrinsic moral value than nonhumans.

Intrinsic Value

Value that an entity or object possesses for its own sake, distinct from its relationship to or utility for any other object.

Extrinsic Value

Value that an entity or object possesses for the sake of something else or in relation to another object, such as for survival, pleasure, or monetary gain.

Strong Anthropocentrism

The philosophical claim that humans are the sole bearers of intrinsic moral value, meaning nonhumans count only extrinsically in relation to humans.

Moderate Anthropocentrism

The philosophical claim that nonhumans possess intrinsic value for their own sake, but significantly less so than humans, meaning human wellbeing prevails in ethical conflicts.

Human-Centred Artificial Intelligence (HCAI)

A paradigm aiming to place humans rather than technology at the centre of AI development, often defined by augmenting human abilities, preserving human control, ensuring safety, and aligning systems with human values.

Asimov's Fourth Law of Robotics

An overarching principle added to Isaac Asimov's robot laws stating that a robot may not harm humanity, or, through inaction, allow humanity to come to harm.

Sentiocentrism

A non-anthropocentric environmental ethic asserting that all sentient animals possessing the capacity to suffer deserve equal moral consideration with humans.

Biocentrism

A life-centred environmental ethic holding that all living entities have intrinsic moral value based on morally significant interests, including survival and flourishing, beyond their utility to humans.

Relational Ecocentric Ethic

An ethical perspective viewing humans, animals, plants, and abiotic systems as interdependent elements of a single interconnected community, where humans are plain members rather than superiors.

Deep Ecology

A branch of ecocentrism that broadens the concept of the self to identify with the wider natural environment, attributing intrinsic value to all components of ecosystems.

Scoping Review

A flexible review methodology designed to provide a comprehensive overview and synthesis of a vast, complex literature base without conducting formal assessments of study bias or quality.

Arksey and O'Malley Five-Step Framework

The methodological framework for scoping reviews comprising: identifying the research question, identifying relevant standards, standard selection, charting data, and collating, summarising, and reporting results.

Trustworthy AI (TAI)

A framework formulated by the European Commission High-Level Expert Group based on four core ethical principles: respect for human autonomy, prevention of harm, fairness, and explicability, unified by human-centredness.

AI Ethics Standards (Non-Technical Tools)

Non-technical instruments—including regulations, professional codes of conduct, standardisation protocols, certification, and governance frameworks—used to implement AI ethics principles.

AI Ethics Guidelines Global Inventory

A public repository developed by AlgorithmWatch that indexed $167$ AI ethics guidelines by May $2022$, utilised as a secondary inventory source in the scoping review.

Indirect Duties Regarding Nature

The Kantian philosophical stance that duties toward nonhuman animals and nature are not owed directly to them, but exist indirectly to avoid degrading human moral character and harming other humans.

BERTlarge Environmental Benchmark

A prominent machine learning language model whose training phase was calculated to generate greenhouse gas emissions comparable to a commercial passenger flight between San Francisco and New York.

GPT-3 Training Environmental Footprint

A $175\text{ billion}$-parameter deep learning model requiring an estimated $1024\text{ GPUs}$ operating for $34\text{ days}$, consuming $936\,\text{MWh}$ of energy at a cost of $\text{USD }4.6\text{ million}$.

Inclusion of Nonhumans (Review Criterion)

A review benchmark met when an AI ethics document explicitly mentions or extends moral, practical, or operational consideration to nonhuman animals, ecology, or the natural environment anywhere within its text.