Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Large language models in climate and sustainability policy: limits and opportunities

T0 review · 4 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read LLM-based text pipelines can classify and summarize climate and sustainability policy documents, but the paper argues they succeed only with substantial human oversight and still lack robust evaluation methods.

desk verdict A genuinely useful workflow paper whose one headline number (80% SDG match) is under-specified; worth peer review, but don't cite the number as it stands. read the letter →

arxiv 2502.02191 v1 pith:HB4AS4SW submitted 2025-02-04 cs.CY

Francesca Larosa , Sergio Hoyas , H. Alberto Conejero , Javier Garcia-Martinez , Francesco Fuso Nerini , Ricardo Vinuesa This is my paper · ORCID
classification cs.CY
keywords largelanguagemodelsclimatepolicysustainabledevelopmentgoalsnaturalprocessingNationallyDeterminedContributionsSDGinterlinkagesworkflowshumanoversight
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to establish that large language models can already perform useful work on climate and sustainability policy texts, processing, classifying, and summarizing documents that are heterogeneous in format and purpose, and that the main barrier to using them in policy is no longer raw capability but human oversight, reproducibility, and evaluation. It supports this with two pilot workflows: one maps paragraphs of Nationally Determined Contributions to the Sustainable Development Goals and classifies their sentiment toward climate action, and the other detects synergies, trade-offs, and directionality among SDGs across a broad corpus of policy and research documents. The strongest quantitative evidence is a reported 80 percent agreement between the LLM routine and expert-assigned SDGs in the second pilot. A sympathetic reader would care because the paper turns the debate away from whether AI can read policy documents toward the practical conditions under which its outputs become usable for policymakers.

What carries the argument

The machinery that carries the argument is the two-pilot workflow itself. The first pilot uses a static semantic-similarity step to assign each NDC paragraph to the SDG with the highest score against UN SDG Fast Facts, then passes the paragraph to ClimateBERT, a climate-domain-adapted model, for sentiment classification on a 0-1-2 scale covering negative, neutral, and positive stances on climate adaptation and mitigation. The second pilot uses a multi-stage prompt on the general-purpose model Gemini 1.0 that first lists main and secondary SDGs for each text and then asks the model to label each SDG pair as synergy, trade-off, or neutral, with directionality stated as inward, outward, or both. The paper's contribution is not a new model but a documented, replicable routine built from these components, plus an explicit account of where the routine fails.

What would settle it

A concrete test would be to run the paper's SDG-assignment prompt on a fresh set of policy documents that have been independently labelled by a panel of experts who did not help build the prompt, and compute agreement on a document-by-document basis; if agreement on unseen documents falls to chance levels, the central claim about LLM usefulness for policy analysis would be undercut.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that a structured LLM-based workflow can reproduce a meaningful portion of expert judgement in sustainability-policy text analysis. When the authors compared their multi-stage prompt routine with the KnowSDGs database of expert-assigned SDG targets, approximately 80 percent of the predicted SDGs matched the experts' assignment, although SDG17 was chronically underdetected and SDG1 appeared in more excerpts than expected. In the NDC pilot, the authors found that combining semantic similarity with UN SDG Fast Facts and ClimateBERT sentiment classification yields a workable paragraph-level mapping of climate pledges to SDGs and to positive, neutral, or negative stances on climate adaptation and mitigation. The paper also documents that the choice of prompt wording changes results for some models, that Gemini 1.0 was selected because its outputs were consistent across runs, and that no labelled benchmark exists to formally evaluate these tasks. The conclusion is that LLMs are useful for nowcasting, large-scale monitoring, and discovering non-obvious SDG connections, but that human intervention at data collection, model selection, and sense-making stages remains necessary.

Load-bearing premise

The paper's load-bearing premise is that the two pilot corpora and the manual screening used to judge outputs are representative of climate and sustainability policy texts generally, and that an LLM's run-to-run consistency is a meaningful sign of quality.

Editorial extensions

If this is right

  • If the results hold, LLM-based routines can act as a reliable first-pass screening tool for NDC and SDG-related policy documents, freeing analysts to focus manual coding on flagged paragraphs.
  • The approximate 80 percent match with expert assignment implies that AI-assisted SDG mapping can complement, but not replace, expert judgement in sustainability policy analysis.
  • The documented sensitivity to prompt wording and model choice implies that published results in this area need to include exact prompts and model versions to be reproducible.
  • The absence of labelled benchmarks means that progress in computational policy analysis will stay difficult to compare until evaluation datasets are built.
  • The workflow demonstrates that grey literature, news items, and national reports can be brought into SDG interlinkage analysis without relying solely on expert-selected academic sources.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the paper used the KnowSDGs database to test successive prompt attempts, the reported 80 percent match may overstate how well the routine would do on genuinely unseen documents; a held-out evaluation against independent expert labels would be a stronger test.
  • Selecting Gemini 1.0 because it is consistent across runs treats determinism as a proxy for quality, which can hide systematic errors; comparing deterministic and stochastic models on the same labelled set would separate stability from accuracy.
  • The trial-and-error prompt process is evidence that prompt wording is becoming a new researcher degree of freedom in computational social science, much like coding schemes were in manual content analysis.
  • A direct extension would be to rerun the same two-pilot workflow on newer NDC submissions or on local-level policy documents to test whether the 80 percent match and the underdetection of SDG17 replicate outside the original corpus.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper reports on two LLM-based text analysis workflows for climate and sustainability policy documents. The first workflow (Section 3.1) processes Nationally Determined Contributions (NDCs) by splitting them into paragraphs, assigning SDGs via semantic similarity against UN SDG Fast Facts, and applying ClimateBERT for sentiment classification; a parallel prompt-based strategy using Gemini 1.0 is also described. The second workflow (Section 3.2) uses a multi-stage prompt on Gemini to assign SDGs to documents and identify synergies, trade-offs, and directionality between SDG pairs, with a comparison against the KnowSDGs database, where the authors report 'approximately 80 percent' agreement. The paper frames these as lessons-learned from two pilot studies, emphasizing the need for human intervention, and discusses limitations grouped by workflow stage (data collection, processing, model selection, evaluation, and translation to policy). The central positive claim—that LLMs are 'successful' at processing, classifying, and summarizing heterogeneous policy text—rests on the 80% figure and on manual screening, neither of which is accompanied by a detailed evaluation protocol.

Significance. The paper addresses a timely and practically important question: whether LLMs can support climate and sustainability policy analysis in a reliable and usable way. Its main strengths are the honest documentation of two concrete workflows—including prompt-engineering iterations, text-cleaning rules, and explicit acknowledgment of evaluation gaps—and the discussion of workflow-stage-specific barriers such as model retirement, market-driven model selection, and the useful-to-usable translation problem. The paper also productively advocates for transparent output evaluation and for hybrid human-machine processes. If the reported 80% agreement with KnowSDGs were backed by a clear evaluation protocol, the paper would provide a useful benchmark for future work. As it stands, the quantitative evidence is too thinly specified to independently validate the abstract's positive claim; the value of the paper is therefore mainly qualitative and methodological, which is still a legitimate contribution if the empirical claims are appropriately qualified.

major comments (4)
  1. [Section 3.2] The paper reports that 'approximately 80 percent of the predicted SDGs match with the experts' assignation' when comparing the LLM routine against KnowSDGs, but it does not define the matching criterion (paragraph-level, document-level, top-k, or exact-match), the denominator, the sample size, or any confidence interval. Since KnowSDGs assigns up to two SDG targets per document while the LLM assigns multiple SDGs per paragraph, the comparison is confounded by unit mismatch; without a precise protocol and ideally precision/recall/F1, this single quantitative claim cannot be independently verified and cannot support the abstract's 'successful' assessment.
  2. [Section 3.1] The selection of Gemini 1.0 over GPT3.5 is justified by the statement that Gemini 'maintains consistency across all runs and is not subject to randomness.' This equates output consistency with model quality, but consistency does not imply accuracy, and no comparison of the two models' SDG assignments against a labeled benchmark is provided. The paper also does not report the decoding parameters (e.g., temperature, top-p, seed) or the number of runs used to establish this consistency, so the model-choice rationale is not reproducible.
  3. [Section 3.1, Section 4] The NDC sentiment and SDG classification results were evaluated by manual screening, which the authors acknowledge 'has its limits' and that 'automatic evaluation is not feasible because of the lack of benchmark datasets.' However, the abstract and Section 4 rely on these manual results to assert that LLMs are 'successful' at classifying and summarizing texts. No inter-annotator agreement statistics, screening sample size, or criteria for resolving disagreements are given, so the central positive claim cannot be independently assessed.
  4. [Section 3.2, Table 3] The multi-stage prompt that is claimed to be a 'fully replicable alternative' is described only at the prompt-text level; the paper does not specify how the output of the first prompt (a variable-length list of main and secondary SDGs) is fed into the second prompt, how the relationship extraction is structured, or how the directionality labels are aggregated. This omission is load-bearing because the central contribution includes the claim of a replicable AI-based routine for detecting synergies and trade-offs; without this pipeline detail, the workflow cannot be re-implemented.
minor comments (5)
  1. [Section 2] The sentence 'Transformers understand logical intra-sequence relationships between different parts and produce outputs based on those... trained on a predefined large corpus on a predefined large corpus of diverse texts' contains a duplicated phrase ('on a predefined large corpus on a predefined large corpus'); please fix this typo.
  2. [Table 3, second prompt] In the trade-off definition, the example 'Climate change will increase poverty by 64%' is labeled 'An example of a synergic relationship,' which is a copy-paste error from the synergy example and could mislead readers implementing the prompt; the label should be corrected.
  3. [Section 3.1] The paper states that Gemini 1.0 'maintains consistency across all runs and is not subject to randomness,' but no decoding parameters (temperature, top-p, random seed) or API version details are reported; please add these for reproducibility.
  4. [Section 4, Figure 4] The paper describes the workflow in Figure 4 but does not provide a data or code availability statement. Since Section 3 claims a 'fully replicable alternative' to expert-based coding, the authors should state whether cleaned corpora, prompts, and analysis scripts will be released, even in a repository.
  5. [References] Reference 18 is cited both for TF-IDF and for the 'attention mechanisms' that introduced the transformer architecture, but the reference list does not include the original Transformer paper (Vaswani et al., 2017); the citations should be corrected and the appropriate reference added.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the paper's empirical LLM outputs are compared against external expert-coded benchmarks, and self-citations are contextual rather than load-bearing.

full rationale

The paper is an empirical workflow description rather than a derivation, and its central claims are checked against external benchmarks. In Section 3.2, the authors report that 'approximately 80 percent of the 'predicted' (i.e., detected by the AI routine) SDGs match with the experts' assignation' when testing against the KnowSDGs database, which is an independently maintained expert-coded resource from the Joint Research Center. This is an external comparison, not a fitted parameter renamed as a prediction. In Section 3.1, SDG labels for NDC paragraphs are obtained through semantic similarity to UN SDG Fast Facts, again an external standard, while sentiment labels are manually screened by the authors with acknowledged limitations. No equation or construction makes the output equal to the input. Self-citations appear in background and motivation sections (e.g., references 2, 7, 15, 34, 44, 58) but they do not carry the central empirical claim; for example, reference 34 is explicitly used as a point of departure ('Rather than surveying experts to check alignment points with the SDGs as in Fuso Nerini et al.'). The acknowledged lack of benchmark datasets and the opaque evaluation protocol for the 80% figure are legitimate transparency and correctness concerns, but they do not constitute circularity-by-construction. The paper's limitations are stated openly rather than hidden, and its main contribution is an empirical application whose success or failure is, in principle, falsifiable against external expert knowledge.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The paper introduces no fitted parameters and no new entities. Its central claims rest on domain assumptions about the validity of semantic similarity as SDG assignment, ClimateBERT sentiment labels, the KnowSDGs ground truth, and the representativeness of the Delphi-selected corpus. These are reasonable defaults but are not independently validated in the paper, and they are the main sources of uncertainty.

assumptions (5)
  • domain assumption Semantic similarity between each NDC paragraph and UN SDG Fast Facts text is a valid proxy for assigning paragraphs to SDGs.
    Used in Section 3.1 first strategy; no comparison against expert-coded labels for the NDC corpus is reported.
  • domain assumption ClimateBERT sentiment categories align with the paper's definitions of 0, 1, and 2 for climate adaptation and mitigation.
    Invoked in Section 3.1 when passing SDG-tagged paragraphs into ClimateBERT; no validation of the mapping is provided.
  • domain assumption The KnowSDGs expert-assigned SDG pair labels are a reliable ground truth for evaluating LLM predictions.
    Section 3.2 uses KnowSDGs as the most appropriate baseline; the paper does not audit the database for bias or coverage gaps.
  • domain assumption Gemini 1.0's consistency across runs implies it is more suitable for the study than GPT3.5.
    Section 3.1; consistency could stem from decoding settings such as temperature, not from better semantic understanding.
  • domain assumption The Delphi method, with the paper's own authors as experts, yields a document corpus representative enough to support general conclusions about SDG interlinkages.
    Section 3.2; self-selection of documents by the authors may bias the corpus and the subsequent analysis.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Large language models in climate and sustainability policy: limits and opportunities." pith.science (2026). https://pith.science/paper/HB4AS4SW

@misc{pith2026250202191,
  author       = {Pith},
  title        = {Pith review of: Large language models in climate and sustainability policy: limits and opportunities},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/HB4AS4SW}},
  note         = {Machine review of arXiv:2502.02191}
}
read the original abstract

As multiple crises threaten the sustainability of our societies and pose at risk the planetary boundaries, complex challenges require timely, updated, and usable information. Natural-language processing (NLP) tools enhance and expand data collection and processing and knowledge utilization capabilities to support the definition of an inclusive, sustainable future. In this work, we apply different NLP techniques, tools and approaches to climate and sustainability documents to derive policy-relevant and actionable measures. We focus on general and domain-specific large language models (LLMs) using a combination of static and prompt-based methods. We find that the use of LLMs is successful at processing, classifying and summarizing heterogeneous text-based data. However, we also encounter challenges related to human intervention across different workflow stages and knowledge utilization for policy processes. Our work presents a critical but empirically grounded application of LLMs to complex policy problems and suggests avenues to further expand Artificial Intelligence-powered computational social sciences.

Figures

Figures reproduced from arXiv: 2502.02191 by the authors.

Figure 1
Figure 1. Comparison of “old” (pre-merged, without adding SpaCy routine) and “new” (post-merged) distributions. a) Paragraphs with a limited number of words are more frequent, but after merging the paragraphs become more comprehensive and move beyond the individual sentence length; b) An example: Andorra’s NDC. Pre-merge, the NDC had 1358 paragraphs; after the merge, Andorra contains 221 paragraphs, 58 words long on average. … view at source ↗
Figure 2
Figure 2. The sentiment and SDG allocation of the first strategy. SDGs are detected using semantic similarity between NDCs’ [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. The geospatial distribution of sentiments and SDGs. Z-score of sentiment indicates the overall tone of countries in describing their actions with respect to climate adaptation and mitigation (upper part); SDGs are grouped according to their core interest (bottom-part). On the environment category: SDG6, SDG13, SDG14 and SDG15; in the society category: SDG1, SDG2, SDG3, SDG4, SDG5, SDG7, SDG11, SDG16; within the econ… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: The application of LLMs to sustainability and climate policy documents in a four [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Exploring the interplay between Planetary Boundaries and Sustainable Development Goals using Large Language Models

    cs.CY 2025-09 reject novelty 5.0 of 10

    Large language models classify over 40,000 climate papers to map synergies and trade-offs between Sustainable Development Goals and Planetary Boundaries.

Reference graph

Works this paper leans on

58 extracted references · 51 canonical work pages · cited by 1 Pith paper

  1. [1]

    Søgaard Jørgensen, P. et al. Evolution of the polycrisis: Anthropocene traps that challenge global sustainability. Philos. Trans. R. Soc. B Biol. Sci. 379, (2024)

    work page 2024

  2. [2]

    Vinuesa, R. et al. The role of artificial intelligence in achieving the Sustainable Development Goals. Nat. Commun. 11, 233 (2020)

    2020

  3. [3]

    & Jung, T

    Koldunov, N. & Jung, T. Local climate services for all, courtesy of large language models. Commun. Earth Environ. 5, 13 (2024)

    work page 2024

  4. [4]

    Wang, H. et al. Scientific discovery in the age of artificial intelligence. Nature 620, 47–60 (2023)

    2023

  5. [5]

    & Malekpour, S

    Bennich, T., Persson, Å., Beaussart, R., Allen, C. & Malekpour, S. Recurring patterns of SDG interlinkages and how they can advance the 2030 Agenda. One Earth 6, 1465–1476 (2023)

    work page 2023

  6. [6]

    & Sompolska-Rzechuła, A

    Kuc-Czarnecka, M., Markowicz, I. & Sompolska-Rzechuła, A. SDGs implementation, their synergies, and trade-offs in EU countries – Sensitivity analysis-based approach. Ecol. Indic. 146, 109888 (2023)

    work page 2023

  7. [7]

    Fuso Nerini, F. et al. Mapping synergies and trade-offs between energy and the Sustainable Development Goals. Nat. Energy 3, 10–15 (2018)

    work page 2018

  8. [8]

    Lam, S. et al. How are climate actions evaluated? A review of United Nations food security evaluations. Glob. Food Sec. 28, (2021)

    work page 2021

Show all 58 references
  1. [9]

    L., Vincent, K., Nicholls, R

    Tompkins, E. L., Vincent, K., Nicholls, R. J. & Suckall, N. Documenting the state of adaptation for the global stocktake of the Paris Agreement. Wiley Interdiscip. Rev. Clim. Chang. 9, (2018)

    2018

  2. [10]

    How needs-based assessments could advance equity in the global stocktake and beyond

    Klinsky, S. How needs-based assessments could advance equity in the global stocktake and beyond. Nat. Clim. Chang. 13, 1007–1009 (2023)

    2023

  3. [11]

    Alcamo, J. et al. Analysing interactions among the sustainable development goals: findings and emerging issues from local and global studies. Sustain. Sci. 15, 1561–1572 (2020)

    2020

  4. [12]

    & Jodlbauer, H

    Tripathi, S., Bachmann, N., Brunner, M., Rizk, Z. & Jodlbauer, H. Assessing the current landscape of AI and sustainability literature: identifying key trends, addressing gaps and challenges. J. Big Data 11, 65 (2024)

    2024

  5. [13]

    & Gärdebo, J

    Wickberg, A. & Gärdebo, J. Computation, data and AI in Anthropocene history. Hist. Technol. 39, 328–346 (2023)

    2023

  6. [14]

    & Crockett, M

    Messeri, L. & Crockett, M. J. Artificial intelligence and illusions of understanding in scientific research. Nature 627, 49–58 (2024)

    2024

  7. [15]

    & Sirmacek, B

    Vinuesa, R. & Sirmacek, B. Interpretable Deep-Learning Models to Help Achieve the Sustainable Development Goals. Interpret. Deep. Model. to Help achieve Sustain. Dev. Goals (2023)

    2023

  8. [16]

    M., Gebru, T., McMillan-Major, A

    Bender, E. M., Gebru, T., McMillan-Major, A. & Shmitchell, S. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? ? in Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency 610–623 (Association for Computing Machinery, 2021). do...

    2021

  9. [17]

    How to shrink AI’s ballooning carbon footprint

    Gibney, E. How to shrink AI’s ballooning carbon footprint. Nature 607, 648 (2022)

    2022

  10. [18]

    & Foege, J

    Maibaum, F., Kriebel, J. & Foege, J. N. Selecting textual analysis tools to classify sustainability information in corporate reporting. Decis. Support Syst. 183, 114269 (2024)

    2024

  11. [19]

    W., Lee, K

    Devlin, J., Chang, M. W., Lee, K. & Toutanova, K. BERT: Pre-training of deep bidirectional transformers for language understanding. NAACL HLT 2019 - 2019 Conf. North Am. Chapter Assoc. Comput. Linguist. Hum. Lang. Technol. - Proc. Conf. 1, 4171–4186 (2019)

    2019

  12. [20]

    & van Rooyen, C

    Stewart, R., Dayal, H., Langer, L. & van Rooyen, C. Transforming evidence for policy: do we have the evidence generation house in order? Humanit. Soc. Sci. Commun. 9, 116 (2022)

    2022

  13. [21]

    H., Winkel, G

    Leipold, S., Feindt, P. H., Winkel, G. & Keller, R. Discourse analysis of environmental policy revisited: traditions, trends, perspectives. J. Environ. Policy Plan. 21, 445–463 (2019)

    2019

  14. [22]

    The SAGE Handbook of Qualitative Data Analysis

    Flick, U. The SAGE Handbook of Qualitative Data Analysis. at https://doi.org/10.4135/9781446282243 (2014)

    2014 doi

  15. [23]

    & Usuelli, M

    Ciscato, E. & Usuelli, M. Words Count: The Role of Language in Overcoming Climate Inertia BT - Interdisciplinary Approaches to Climate Change for Sustainable Growth. in (eds. Valaguzza, S. & Hughes, M. A.) 27–42 (Springer International Publishing, 2022). doi:10.1007/978 -3-030...

    2022 doi

  16. [24]

    & Guenther, L

    Mahl, D. & Guenther, L. Content Analysis in the Research Field of Environmental & Climate Change Coverage. in Oehmer-Pedrazzi, F., Kessler, S.H., Humprecht, E., Sommer, K., Castro, L. (eds) (eds. Oehmer-Pedrazzi, F., Kessler, S. H., Humprecht, E., Sommer, K. & Castro, L.) 203–...

    2023 doi

  17. [25]

    & Allan, B

    Meckling, J. & Allan, B. B. The evolution of ideas in global climate policy. Nat. Clim. Chang. 10, 434–438 (2020)

    2020

  18. [26]

    & Wilson, S

    Fleming, A., Vanclay, F., Hiller, C. & Wilson, S. Challenging dominant discourses of climate change. Clim. Change 127, 407–418 (2014)

    2014

  19. [27]

    W., Minx, J

    Callaghan, M. W., Minx, J. C. & Forster, P. M. A topography of climate change research. Nat. Clim. Chang. 10, 118–123 (2020)

    2020

  20. [28]

    Falkenberg, M. et al. Growing polarization around climate change on social media. Nat. Clim. Chang. 12, 1114–1121 (2022)

    2022

  21. [29]

    & Boykoff, M

    McAllister, L., Vedula, S., Pu, W. & Boykoff, M. Vulnerable voices: using topic modeling to analyze newspaper coverage of climate change in 26 non-Annex I countries (2010-2020). Environ. Res. Lett. 19, (2024)

    2024

  22. [30]

    & Palacios, H

    Luccioni, A. & Palacios, H. Using Natural Language Processing to Analyze Financial Climate Disclosures. Lr. 2018 - 11th Int. Conf. Lang. Resour. Eval. 4470–4473 (2019)

    2019

  23. [31]

    Schäfer, M. S. & Hase, V. Computational methods for the analysis of climate change communication: Towards an integrative and reflexive approach. Wiley Interdiscip. Rev. Clim. Chang. 14, 1–10 (2023)

    2023

  24. [32]

    Cooperative equilibrium of the China-US-EU climate game

    Da Zhu, J. Cooperative equilibrium of the China-US-EU climate game. Energy Strateg. Rev. 39, 100797 (2022)

    2022

  25. [33]

    Nationally determined contributions (NDCs) as a governance instrument – accounting for politics, negotiation progress, and related mechanisms under the Paris Agreement

    Leiter, T. Nationally determined contributions (NDCs) as a governance instrument – accounting for politics, negotiation progress, and related mechanisms under the Paris Agreement. Env. Polit. 33, 552–557 (2024)

    2024

  26. [34]

    Nerini, F. et al. Connecting Climate Action with other Sustainable Development Goals. Nat. Sustain. 2, 674–680 (2019)

    2019

  27. [35]

    A., Leippold, M

    Webersinke, N., Kraus, M., Bingler, J. A., Leippold, M. & Erlangen-nuremberg, F. A. U. ClimateBERT: A Pretrained Language Model for Climate-Related Text. arXiv:2110.12010 (2022)

    2022 arXiv

  28. [36]

    Gemini : A Family of Highly Capable Multimodal Models

    Gemini Team, G. Gemini : A Family of Highly Capable Multimodal Models. arXiv Comput. Sci. 1–90 (2024)

    2024

  29. [37]

    Our Common Future (‘The Brundtland Report’)

    World Commission on Environment and Development. Our Common Future (‘The Brundtland Report’). United Nations 52–55 (1987) doi:10.4324/9781351279086-15

    1987 doi

  30. [38]

    A threefold approach to rescue the 2030 Agenda from fa iling

    Pradhan, P. A threefold approach to rescue the 2030 Agenda from fa iling. Natl. Sci. Rev. 10, nwad015 (2023)

    2023

  31. [39]

    Carmona, R. et al. Analysing engagement with Indigenous Peoples in the Intergovernmental Panel on Climate Change’s Sixth Assessment Report. npj Clim. Action 2, 29 (2023)

    2023

  32. [40]

    Keep it complex

    Stirling, A. Keep it complex. Nature 468, 1029–1031 (2010)

    2010

  33. [41]

    Cash, D. W. et al. Knowledge systems for sustainable development. Proc. Natl. Acad. Sci. 100, 8086–8091 (2003)

    2003

  34. [42]

    Evolving the IPCC

    Nature Climate Change Editor. Evolving the IPCC. Nat. Clim. Chang. 1, 227 (2011)

    2011

  35. [43]

    J., Le Quéré, C., Smith, A

    De-Gol, A. J., Le Quéré, C., Smith, A. J. P. & Aubin Le Quéré, M. Broadening scientific engagement and inclusivity in IPCC reports through collaborative technology platforms. npj Clim. Action 2, 49 (2023)

    2023

  36. [44]

    Larosa, F. et al. Halting generative AI advancements may slow down progress in climate research. Nat. Clim. Chang. 13, 497–499 (2023)

    2023

  37. [45]

    Arnott, J. C. & Lemos, M. C. Understanding knowledge use for sustainability. Environ. Sci. Policy 120, 222–230 (2021)

    2021

  38. [46]

    J., Ford, J

    Sietsma, A. J., Ford, J. D., Callaghan, M. W. & Minx, J. C. Progress in climate change adaptation research. Environ. Res. Lett. 16, 54038 (2021)

    2021

  39. [47]

    Callaghan, M. et al. Machine-learning-based evidence and attribution mapping of 100,000 climate impact studies. Nat. Clim. Chang. 11, 966–972 (2021)

    2021

  40. [48]

    Singhal, K. et al. Large language models encode clinical knowledge. Nature 620, 172–180 (2023)

    2023

  41. [49]

    Martini, G. et al. Machine learning can guide food security efforts when primary data are not available. Nat. Food 3, 716–728 (2022)

    2022

  42. [50]

    E., Tzachor, A., Avin, S

    Richards, C. E., Tzachor, A., Avin, S. & Fenner, R. Rewards, risks and responsible deployment of artificial intelligence in water systems. Nat. Water 1, 422–432 (2023)

    2023

  43. [51]

    Evaluating large language models for use in healthcare: A framework for translational value assessment

    Reddy, S. Evaluating large language models for use in healthcare: A framework for translational value assessment. Informatics Med. Unlocked 41, 101304 (2023)

    2023

  44. [52]

    Ullah, E., Parwani, A., Baig, M. M. & Singh, R. Challenges and barriers of using large language models (LLM) such as ChatGPT for diagnostic medicine with a focus on digital patholog y - a recent scoping review. Diagn. Pathol. 19, 43 (2024)

    2024

  45. [53]

    Zhou, C. et al. A Comprehensive Survey on Pretrained Foundation Models: A History from BERT to ChatGPT. 1–99 (2023)

    2023

  46. [54]

    Liu, P. et al. Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing. ACM Comput. Surv. 55, 1–46 (2023)

    2023

  47. [55]

    Ai Language Models Technological, Socio-Economic and Policy Considerations

    OECD. Ai Language Models Technological, Socio-Economic and Policy Considerations. Oecd Digit. Econ. Pap. 352, 1–1 (2023)

    2023

  48. [56]

    Crochemore, L. et al. A Framework for Joint Verification and Evaluation of Seasonal Climate Services across Socioeconomic Sectors. Bull. Am. Meteorol. Soc. 105, E1218–E1236 (2024)

    2024

  49. [57]

    Calliari E., S. S. & Mysiak, J. The Politics of (and behind) the UNFCCC’s Loss and Damage Mechanism. in Loss and Damage from Climate Change. Concepts, Methods and Policy Options (ed. Mechler, R. B. L. S. T. S. S. L.-B. J.) 155–178 (Springer, 2019)

    2019

  50. [58]

    ERDF A Way of Making Europe

    Larosa, F. & Wickberg, A. Artificial Intelligence can help Loss and Damage only if it is inclusive and accessible. npj Clim. Action 3, 59 (2024). Competing interests The authors declare no competing interests, financial or other, exist. Funding statement RV, FFN and FL acknowl...

    2024

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.