pith. sign in

arxiv: 2604.15327 · v1 · submitted 2026-03-06 · 💻 cs.HC · cs.AI

Eco-Bee: A Personalised Multi-Modal Agent for Advancing Student Climate Awareness and Sustainable Behaviour in Campus Ecosystems

Caleb Adu , Neil Kapadia , Binhe Liu , Jonathan Randall , Sruthi Viswanathan This is my paper

Pith reviewed 2026-05-15 15:36 UTC · model grok-4.3

classification 💻 cs.HC cs.AI
keywords Eco-Beeplanetary boundariessustainable behaviourAI conversational agentcampus ecosystemsclimate awarenessbehavioural changeEco-Score
0
0 comments X

The pith

Eco-Bee uses an AI conversational agent to map students' daily choices onto planetary boundaries through an Eco-Score and gamified feedback.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper proposes Eco-Bee as a multi-modal agent that combines large language models with a translation of the Planetary Boundaries framework into a personal Eco-Score. The agent supplies students with actionable insights, peer comparisons, and challenges that link routine campus behaviors in food, transport, and energy to global environmental limits. A short pilot across campus networks with 52 participants found that 96 percent supported a full rollout and reported improved grasp of how individual actions add up to collective planetary pressure. The work positions this integration as a step beyond static carbon calculators toward sustained behavior change in university settings.

Core claim

Eco-Bee integrates large language models, a translation of the Planetary Boundaries framework as Eco-Score, and a conversational agent that connects individual choices to environmental limits, delivering actionable insights, peer benchmarking, and gamified challenges to sustain engagement and drive measurable progress toward boundary-aligned living.

What carries the argument

Eco-Bee, the multi-modal conversational agent that personalizes advice by converting daily behaviors into an Eco-Score derived from planetary boundaries.

Load-bearing premise

Self-reported clarity and rollout support from a small, short-term pilot will translate into measurable and sustained changes in actual student behavior.

What would settle it

A controlled follow-up study that tracks objective metrics such as campus resource use or verified emissions reductions among Eco-Bee users versus a non-user group over at least one academic year.

Figures

Figures reproduced from arXiv: 2604.15327 by Binhe Liu, Caleb Adu, Jonathan Randall, Neil Kapadia, Sruthi Viswanathan.

Figure 1
Figure 1. Figure 1: Overview of Eco-Bee: (Left) initial idea used for co-design interactions; (Centre) screenshot [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
read the original abstract

Universities are microcosms of urban ecosystems, with concentrated consumption patterns in food, transport, energy, and product usage. These environments not only contribute substantially to sustainability pressures but also provide a unique opportunity to advance sustainability education and behavioural change at scale. As in most sectors, digital sustainability initiatives within universities remain narrowly focused on carbon calculations, typically providing static feedback that limits opportunities for sustained behavioural change. To address this gap, we propose Eco-Bee, integrating large language models, a translation of the Planetary Boundaries framework (as Eco-Score), and a conversational agent that connects individual choices to environmental limits. Tailored for students at the cusp of lifelong habits, Eco-Bee delivers actionable insights, peer benchmarking, and gamified challenges to sustain engagement and drive measurable progress toward boundary-aligned living. In a pilot tested across multiple campus networks (n=52), 96% of the student participants supported a campus-wide rollout and reported a clearer understanding of how daily behaviours collectively impact the planet's limits. By embedding planetary science, behavioural reinforcement, and AI-driven personalisation into a single platform, Eco-Bee establishes a scalable foundation for climate-conscious universities and future AI-mediated sustainability infrastructures.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes Eco-Bee, a multi-modal conversational agent that combines large language models with a translation of the Planetary Boundaries framework (Eco-Score) to deliver personalized feedback, peer benchmarking, and gamified challenges aimed at improving university students' climate awareness and sustainable behaviors in campus ecosystems. It reports a pilot deployment across campus networks (n=52) in which 96% of participants supported a campus-wide rollout and reported clearer understanding of how daily behaviors impact planetary limits.

Significance. If the effectiveness claims are substantiated, Eco-Bee could offer a scalable model for embedding planetary-boundaries science into everyday student decision-making via AI personalization and behavioral reinforcement, moving beyond static carbon calculators common in current university sustainability tools.

major comments (2)
  1. [Pilot evaluation] Pilot evaluation (described in the results section): the reported 96% support and improved understanding rest entirely on self-reported survey responses from n=52 participants with no control arm, no pre/post objective behavioral proxies (energy use, waste audits, transport logs), and no longitudinal follow-up. This design cannot support the central claim of driving measurable progress toward boundary-aligned living.
  2. [System architecture] Eco-Score definition (system architecture section): the mapping from individual actions to specific Planetary Boundaries is presented as a direct translation, yet the manuscript provides no explicit algorithm, weighting scheme, or validation against empirical boundary data, leaving the score's reliability and parameter independence unclear.
minor comments (2)
  1. [Abstract] Abstract: the phrase 'measurable progress' should be qualified given that the pilot supplies only self-reported understanding rather than objective metrics.
  2. [Figures and tables] Figure captions and tables: ensure all pilot survey items, response scales, and demographic breakdowns are fully reported so readers can assess selection bias and generalizability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for these constructive comments, which help clarify the scope and limitations of our pilot work. We address each point below and will revise the manuscript accordingly to strengthen transparency and align claims with the available evidence.

read point-by-point responses

  1. Referee: [Pilot evaluation] Pilot evaluation (described in the results section): the reported 96% support and improved understanding rest entirely on self-reported survey responses from n=52 participants with no control arm, no pre/post objective behavioral proxies (energy use, waste audits, transport logs), and no longitudinal follow-up. This design cannot support the central claim of driving measurable progress toward boundary-aligned living.

    Authors: We agree that the pilot relies exclusively on self-reported data from a modest sample and lacks controls, objective proxies, or follow-up, which prevents strong causal claims about measurable behavioral change. The abstract's reference to 'measurable progress' is aspirational for the overall Eco-Bee vision rather than a direct claim about the pilot results, which focus on acceptance and perceived understanding. We will revise the results, discussion, and abstract to explicitly frame the study as preliminary, add a limitations subsection detailing these design constraints, and outline concrete plans for future controlled trials that incorporate objective metrics such as energy-use logs and waste audits. revision: yes

  2. Referee: [System architecture] Eco-Score definition (system architecture section): the mapping from individual actions to specific Planetary Boundaries is presented as a direct translation, yet the manuscript provides no explicit algorithm, weighting scheme, or validation against empirical boundary data, leaving the score's reliability and parameter independence unclear.

    Authors: The Eco-Score is presented as a simplified, user-facing approximation of Planetary Boundaries rather than a fully validated quantitative model. We will expand the system architecture section with an explicit description of the action-to-boundary mapping algorithm, the literature-derived weighting scheme, and a clear discussion of its simplifications, absence of comprehensive empirical validation, and assumptions regarding parameter independence. These additions will improve reproducibility and allow readers to evaluate the score's reliability. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper presents Eco-Bee as an integration of large language models with a direct translation of the external Planetary Boundaries framework into Eco-Score, plus a conversational agent for personalization and gamification. No mathematical derivations, fitted parameters, or equations appear that reduce any claimed result to its own inputs by construction. The pilot (n=52) reports self-reported support and understanding as empirical observations without predictive modeling or self-referential fitting. Central claims rest on external frameworks and pilot data rather than self-citation chains or ansatzes that collapse into the paper's own definitions.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

The central claim rests on the assumption that Planetary Boundaries can be meaningfully translated into personal Eco-Scores and that conversational AI can drive sustained behavior change; no free parameters or invented physical entities are defined.

axioms (1)
  • domain assumption Planetary Boundaries framework can be translated into an actionable personal Eco-Score for daily choices
    Invoked to connect individual behaviors to global limits without further justification in the abstract.
invented entities (2)
  • Eco-Bee agent no independent evidence
    purpose: Personalized multi-modal conversational agent for climate awareness
    New system proposed by the authors.
  • Eco-Score no independent evidence
    purpose: Personal translation of Planetary Boundaries for feedback
    New metric introduced to operationalize the framework.

pith-pipeline@v0.9.0 · 5527 in / 1332 out tokens · 37436 ms · 2026-05-15T15:36:50.205206+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

9 extracted references · 9 canonical work pages

  1. [1]

    Evidence from two large field experiments that peer comparison feedback can reduce residential energy usage.Journal of Law, Economics, and Organization, 29(5):992–1022, 2013

    Ian Ayres, Sophie Raseman, and Alice Shih. Evidence from two large field experiments that peer comparison feedback can reduce residential energy usage.Journal of Law, Economics, and Organization, 29(5):992–1022, 2013

    work page 2013

  2. [2]

    S, tefan Boncu, Octav-Sorin Candel, and Nicoleta Laura Popa. Gameful green: A systematic review on the use of serious computer games and gamified mobile apps to foster pro-environmental information, attitudes and behaviors.Sustainability, 14(16), 2022

    work page 2022

  3. [3]

    Coolclimate calculator: Estimates greenhouse gas emissions across multiple domains with personalised footprints and action planning

    CoolClimate Network, University of California, Berkeley. Coolclimate calculator: Estimates greenhouse gas emissions across multiple domains with personalised footprints and action planning. Accessed August 2025

    work page 2025

  4. [4]

    Breaking free from tunnel vision for climate change and health.PLOS Global Public Health, 3(3):e0001684, 2023

    Thilagawathi Abi Deivanayagam and Rhiannon Elizabeth Osborne. Breaking free from tunnel vision for climate change and health.PLOS Global Public Health, 3(3):e0001684, 2023

    work page 2023

  5. [5]

    Eric Lewandowski, and Elouise E

    Caroline Hickman, Elizabeth Marks, Panu Pihkala, Susan Clayton, R. Eric Lewandowski, and Elouise E. Mayall. Climate anxiety in children and young people and their beliefs about govern- ment responses to climate change: a global survey.The Lancet Planetary Health, 5(12):e863– e873, 2021

    work page 2021

  6. [6]

    Otto, Jonathan F

    Ilona M. Otto, Jonathan F. Donges, Roger Cremades, Avit Bhowmik, Richard J. Hewitt, Wolfgang Lucht, Johan Rockström, Franziska Allerberger, Mark McCaffrey, Sylvanus S. P. Doe, Alex Lenferna, Nerea Morán, Detlef P. van Vuuren, and Hans Joachim Schellnhuber. Social tipping dynamics for stabilizing earth’s climate by 2050.Proceedings of the National Academy ...

    work page 2050

  7. [7]

    Stuart Chapin, Eric Lambin, Timothy Lenton, Marten Scheffer, Carl Folke, Hans Joachim Schellnhuber, Björn Nykvist, Cynthia A

    Johan Rockström, Will Steffen, Kevin Noone, Åsa Persson, F. Stuart Chapin, Eric Lambin, Timothy Lenton, Marten Scheffer, Carl Folke, Hans Joachim Schellnhuber, Björn Nykvist, Cynthia A. de Wit, Terry Hughes, Sander van der Leeuw, Henning Rodhe, Sverker Sörlin, Peter K. Snyder, Robert Costanza, Uno Svedin, Malin Falkenmark, Louise Karlberg, Robert W. Corel...

    work page 2009

  8. [8]

    How is climate change impacting the educational choices and career plans of undergraduates?Sustainability, 17(14), 2025

    Sarah Lynne Stafford. How is climate change impacting the educational choices and career plans of undergraduates?Sustainability, 17(14), 2025

    work page 2025

  9. [9]

    leaderboard

    Michael P. Vandenbergh and Paul C. Stern. The role of individual and household behaviour in decarbonization.Environmental Law Reporter, 47:10941–10946, 2017. 4 A Pseudocode for Quiz Flow and EcoScore Computation A.1 Quiz Flow (Frontend and API) state Profile = { cohort, campus, faculty, pseudonym, quiz: { diet, mobility, fashion, housing, career_interest,...

    work page 2017