AI assisted tools

Before you use the tools

Treat outputs as structured drafts, hypotheses, and starting points for investigation - not as validated Welfare Footprint estimates or official WFI scientific assessments.
Use human scientific and domain-expert review for consequential interpretation or decisions.
Do not enter confidential, sensitive, personal, or proprietary material. Use public, anonymized, or redacted inputs unless an approved WFI process explicitly permits otherwise.
Check important terminology, methods, and claims against the current Scientific Canon and cited primary sources.
When attributing an output, record the tool name, WFI, the visible model/version if available, the access date, the URL, and that the material is an AI-generated draft.

What these tools are — and are not

These tools are intended to support learning, exploration, and first-pass analyses with the Welfare Footprint Framework (WFF). They do not produce WFI estimates, and they do not represent official scientific assessments by the Welfare Footprint Institute. Published WFI analyses are developed separately through comprehensive evidence review, transparent documentation of assumptions, quality checks, and scientific scrutiny.

Like any AI-assisted tool, the outputs should be treated as structured hypotheses and starting points for further investigation rather than definitive conclusions. Because generative AI is not deterministic, exploring a question through multiple conversations or prompts will often produce a more complete and robust understanding than relying on a single interaction.

How to cite

Cite the Welfare Footprint Framework (WFF) with the public Scientific Canon reference. When you cite an AI tool output, name the tool, the Welfare Footprint Institute, the model or version if it is visible, the access date, and the URL, and label the material “AI-generated draft.” AI tool outputs are not WFI estimates.

Framework

Alonso, W. J., & Schuck-Paim, C. (2025). Welfare Footprint Framework: Methodological Foundations and Quantitative Assessment Guidelines. Center for Welfare Metrics (São Paulo). https://doi.org/10.17605/osf.io/94bxs Download PDF

AI tool output

Hedonic-Track GPT, Welfare Footprint Institute, model and version if visible, accessed October 9, 2026, https://chatgpt.com/g/g-cpXxeWtgg-hedonic-track. AI-generated draft. Not a WFI estimate.

If a figure summarizes several runs, also report the median and the range from low to high.

Which tool should I use?

Question or task	Start with
Understand the WFF or check terminology	Welfare Footprint Framework companion
Describe how animals live or what circumstances change	(i) Zootechnical Mapper GPT
Map Biological Consequences and Affective Experiences	(ii) AffectMap GPT
Draft a Pain-Track, Pleasure-Track, or affective quantification	(iii) Hedonic-Track GPT
Estimate prevalence, exposure, frequency, or population occurrence	(iv) Epidemiology for Animal Welfare GPT
Examine evidence about affective capacity across species	(Ψ) interspecific affect GPT
Stress-test a research idea, proposal, study, or report	Animal welfare research robustness reviewer
Assess an intervention against a counterfactual	Animal welfare impact assessment assistant
Explore conceptual questions about affect and consciousness	Neurophilosopher GPT
Explore a consumer-facing application of welfare analysis	Food welfare explorer

Why the Roman numerals and Ψ matter

The labels (i), (ii), (iii), (iv), and (Ψ) identify the Welfare Footprint Framework module in which each analytical workflow is rooted.

(i) Module I - describe the animals, Life-Fates, phases, environments, management, and other Circumstances.
(ii) Module II - map how Circumstances can lead to Biological Consequences and Affective Experiences.
(iii) Module III - characterize and quantify negative or positive Affective Experiences.
(iv) Module IV - estimate how many animals are affected, how often, and with what uncertainty.
(Ψ) Interspecific analysis - examine welfare-relevant capacities and the assumptions needed for comparisons across species.

WFF module-rooted tools

The examples below reproduce current public conversation starters or repository-documented interaction patterns. They illustrate what to enter into the specific tool, not everything its underlying WFF module could theoretically do. Short inputs are intentional: several tools guide the user through a step-by-step workflow.

(i) Zootechnical Mapper GPT

WFF relationship: Rooted in Module I.

This tool structures the Circumstances animals experience across a life, phase, procedure, production system, wild-animal context, or intervention. It helps describe what changes before inferring welfare consequences.

What can you use it for?

Describe a production system and the Circumstances animals experience without inferring welfare consequences.
Organize information by Life-Fates, Life-Phases, internal and external Circumstances, and Species-Specific Needs.
Compare systems or describe conditions before and after an intervention.
Structure information from reports or papers and identify important gaps.

Examples of questions/prompts

“Layer hen system in the EU”
“Intensive pig system in China”
“Atlantic salmon sea-cage in Norway”

Access: Open the GPT

(ii) AffectMap GPT

WFF relationship: Rooted in Module II.

AffectMap helps organize causal pathways from Circumstances to Biological Consequences and Affective Experiences. It is useful for making direct, indirect, and unintended welfare pathways visible.

What can you use it for?

Map pathways from Circumstances to Biological Consequences and Affective Experiences.
Identify both harmful and beneficial effects.
Surface overlooked pathways, additive interactions and alternative explanations.
Plan what indicators and evidence would be needed to examine the proposed pathways.

Examples of questions/prompts

“Long-term caging with bile extraction procedures (bears)”
“Pastures rotation with shade, clean water points, and comfortable lying areas (dairy cows)”
“Group-housed sows with deep bedding and manipulable enrichment (pigs)”

Access: Open the GPT

(iii) Hedonic-Track GPT

WFF relationship: Rooted in Module III.

Hedonic-Track supports first-pass affective quantification for a specific negative or positive experience. It can draft Pain-Tracks, Pleasure-Tracks, Cumulative Pain, or Cumulative Pleasure analyses for scientific review.

What can you use it for?

Develop a first-draft intensity-over-time trajectory for a specific negative or positive Affective Experience.
Structure Pain-Track or Pleasure-Track temporal segments.
Make evidence, assumptions and uncertainty explicit.
Compare plausible trajectories or scenarios for subsequent scientific review.

Examples of questions/prompts

“Surgical castration in piglets”
“The play of a pig when straw is provided”
“Asphyxia due to air exposure in salmon”

Access: Open the GPT | Example analyses

(iv) Epidemiology for Animal Welfare GPT

WFF relationship: Rooted in Module IV.

This tool helps develop population-level parameters: how many animals are affected, how often an experience or Biological Consequence occurs, how representative the evidence is, and how uncertainty should be represented. It does not quantify the individual affective experience or calculate the final Welfare Footprint.

What can you use it for?

Estimate how common, long-lasting or widely distributed an animal-welfare problem may be.
Review epidemiological evidence across populations, systems, regions and time periods.
Distinguish prevalence, incidence, duration and other measures and denominators.
Identify bias, heterogeneity, evidence gaps and research priorities.

Examples of questions/prompts

“Severe lameness in market pigs in intensive indoor systems (global, current typical conditions)”
“Respiratory disease/pneumonia in nursery pigs in intensive production (Brazil, last 5 years)”
“Footpad dermatitis in broiler chickens in conventional indoor housing (EU, current typical conditions)”

Access: Open the GPT | Open the GPT

(Ψ) Interspecific Affect GPT

WFF relationship: Rooted in the WFF interspecific analysis.

This tool takes one target species or taxon at a time. It first assesses whether sentience is a coherent and plausible target, then assembles relevant behavioral, neural, physiological, pharmacological, cognitive, and evolutionary or ecological evidence. Its default output is a provisional Pain ceiling expressed using qualitative human-anchored reference categories: Annoying(h), Hurtful(h), Disabling(h), and Excruciating(h). These are reference anchors, not equivalence claims. The tool does not directly compare species, rank them, assign moral weights, or calculate Welfare Footprints.

What can you use it for?

Begin with one target species or taxon and proceed through the analysis one step at a time.
Assess sentience plausibility before attempting a ceiling inference.
Evaluate evidence and competing explanations relevant to high-intensity affective capacity.
Develop a provisional Pain-ceiling category or bounded range on the human-anchored (h) scale, with explicit uncertainty. A Pleasure ceiling is considered only when explicitly requested.

Examples of questions/prompts

Enter one target species or taxon. The GPT will guide the analysis one step at a time.

“shrimp”
“tilapia”
“dog”

Access: Open the GPT

Cross-cutting tools and applications

Welfare Footprint Framework Companion

WFF relationship: Cross-cutting framework orientation; not a module GPT.

The Companion helps users navigate WFF concepts, terminology, modules, calculations, and source materials. It can help structure an impact question or review whether a draft uses WFF distinctions consistently.

What can you use it for?

Orient an animal-welfare question within the WFF analytical chain and modules.
Check terminology and methodology against the WFI Scientific Canon available to the tool.
Distinguish evidence from assumptions and identify missing information.
Develop WFF-aligned research, review and impact-assessment plans.

Examples of questions/prompts

“I will provide a passage. Check whether its terminology and methodology align with the WFF.”
“Help me identify where this question fits within the WFF analytical chain and modules.”
“I will provide a welfare claim. Identify what the WFF supports, what remains uncertain, and what evidence is missing.”

Access: Open the GPT

Animal Welfare Research Robustness Reviewer

WFF relationship: Cross-cutting research review.

This reviewer stress-tests research ideas, intervention proposals, study designs, grant concepts, models, reports, and draft analyses. It asks whether the rationale, assumptions, design, feasibility, outcomes, and decision relevance are strong enough for the next stage.

What can you use it for?

Review an early-stage research idea or proposal.
Assess study design, outcome definitions and analysis plans.
Stress-test an intervention evaluation or welfare-estimation model.
Identify unintended effects, overclaims and WFF relevance.

Examples of questions/prompts

Select a starter, then upload or paste the research material to be reviewed.

“Upload proposal to review”
“Assess study design”
“Strengthen this study idea”

Access: Open the GPT

Animal Welfare Impact Assessment Assistant

WFF relationship: Cross-cutting intervention assessment and decision support.

This assistant supports a first-pass comparison of an animal-welfare intervention against a counterfactual. It maps direct and indirect effects, challenges assumptions and evidence, identifies uncertainty, and helps decide when scenario analysis, WFF methods, or specialist review are appropriate.

What can you use it for?

Define an intervention and its relevant counterfactual.
Map direct and indirect effects and the animal populations affected.
Test assumptions, evidence, unintended effects and uncertainty.
Determine whether scenario analysis, quantification, WFF methods or specialist review are appropriate.

Examples of questions/prompts

Select a starter, then describe the intervention, technologies, or proposed impact.

“Assess this welfare intervention”
“Compare two welfare technologies”
“Check whether this can be quantified”

Access: Open the GPT

Neurophilosopher GPT

WFF relationship: Cross-disciplinary conceptual exploration; not a WFF module package.

This tool explores affective states and consciousness through neuroscience and philosophical reasoning. It is intended for conceptual exploration, not for producing a Welfare Footprint estimate.

What can you use it for?

Explore ideas from philosophy, religion and related fields.
Distinguish emotional and intellectual components of a theory or concept.
Relate conceptual claims to neuroscience and affective cognition.
Compare interpretations and expose underlying assumptions.

Examples of questions/prompts

“John Stuart Mill’s concepts of lower and higher pleasures”
“Descartes’ Dualism”
“Plato’s Tripartite Soul”

Access: Open the GPT

Food Welfare Explorer

WFF relationship: Consumer-facing web application; not a Custom GPT or WFF module package.

Food Welfare Explorer is a consumer-facing prototype. Users submit a food item by image or text, confirm or refine the detected animal-derived item, review a welfare-oriented analysis, select an ethical lens, and request suitable alternatives. Its current AI-assisted outputs are exploratory, not definitive product-level Welfare Footprint estimates.

What can you use it for?

Submit a food item by image or text.
Confirm or refine the animal-derived item detected by the application.
Review the welfare-oriented analysis provided for that item.
Select or change an ethical lens and request functionally appropriate alternatives.

Examples of questions/prompts

For this web application, these are actions to try rather than chat prompts:

“Upload a photo of a food item and confirm or refine the animal-derived item detected.”
“Enter a food item by text and review the welfare-oriented analysis.”
“Change the ethical lens, then request alternatives consistent with that lens.”

Access: Open the web application

Using the tools together

Each tool answers a different question. A combined workflow can make an analysis more transparent without implying that every question needs every tool.

1.
Use the Zootechnical Mapper to describe the baseline and alternative Circumstances without yet inferring welfare consequences.
2.
Use AffectMap to map plausible Biological Consequences and Affective Experiences, including indirect and unintended pathways.
3.
Use Hedonic-Track for first-pass quantification of priority negative or positive experiences at the individual level.
4.
Use Epidemiology for Animal Welfare to estimate how many animals are affected, how often, and with what uncertainty.
5.
Use the Companion to check WFF distinctions and the Scientific Canon for authoritative terminology and methodology. Use the Robustness Reviewer or Impact Assessment Assistant to stress-test study design, counterfactuals, assumptions, feasibility, and decision relevance.

Example: comparing urban rat-control methods might begin with the circumstances created by anticoagulant rodenticides, glue traps, and snap traps; map the resulting biological and affective pathways; characterize priority experiences; estimate affected populations; and then examine indirect effects such as population recovery, immigration, disease, hunger, or aggression. The result is a structured research map, not an automatically validated impact estimate.