AI Assisted Paper Reviewing

A Necessary Reform to Fix a Broken System

Apr 18, 2025

Sometimes, the Goddess Gaia is asked to review scientific papers on the ecosystem written by human authors. She is not always happy with what she reads.

Everybody says that the reviewing processing in scientific journals is broken, but nobody seems to be able to find something better. Well; I have an idea: why don’t we try “AI Assisted Paper Reviewing?”

The problem with human reviewers is that they are often lazy and incompetent; in addition, they may be ideologically biased or simply have a personal grudge against one of the authors. Besides, in many cases, reviewers seem to think that any paper proposing some non-standard or innovative idea is heresy and hence should be burned in such a way as to leave nothing but ashes.

The results are disastrous. The reviewing process is a bottleneck that sometimes can’t filter out truly bad papers, while it stifles innovation and encourages mediocrity. In addition, people pushing cranky or ideologically bent concepts find an easy way to justify their lack of success in publishing their work by blaming the reviewing system, biased against them. If you are a working scientist, you know what I am talking about.

So, here is the idea. Right now, the best AIs can perfectly well review a paper on their own. I experimented with “DeepSearch” on two papers: one which I judged pretty bad (see my previous post) and one which I wrote with my colleagues, and which, for obvious reasons, I see as good :-). The AI could correctly pinpoint the mistakes and the misinterpretations of the “bad” paper, while it appreciated the sound approach of the “good” one.

Below, you’ll find the complete report created by “Deepsearch” (the cousin of Grok) in a few minutes. It is a review that I daresay is better than most human reviews I received for my papers. Many of them clearly showed that the reviewer hadn’t bothered to read the paper, or to make even a minimal effort to understand what it was about.

Deepsearch can be criticized for a few points. For instance, it says that the paper lacks “new empirical data,” which is obvious, since it is a review. It also criticizes “speculation” — but it doesn’t have a clear idea of what “speculation” is. Speculation is a hypothesis with no data. A hypothesis is an unproven thesis proposed on the basis of data. It is perfectly legitimate to propose an unproven thesis in a scientific paper; it is the way science advances: by proving unproven theses. But, overall, the review goes in deep into the paper, it shows no biases, and it is very well organized and done. It still needs a “meta-review” by a human expert, but it shows the way to go.

The best AIs are neutral and not ideologically bent; they don’t hate anyone (I hope), and they have a depth of knowledge that no human can match. In addition, they cost very little, are fast, and accurate. So, why don’t we use this capability for a first “filter” of a submission? Then, the AI evaluation can be passed to a human reviewer, who will provide a final evaluation and recommendation.

This is just a test, but if readers have time, they could verify how well AIs can manage reviews of published or papers to be published using different AIs. In any case, AI reviews of scientific papers are one of these unavoidable things that, by definition, are inevitable (also, inevitable things tend to be unavoidable). If journals do not adopt this method of review, reviewers will. Whether AIs will be able to write papers by themselves is all to be seen, but I am afraid it will be soon.

If you would like to have a preprint of our paper, just tell me in the comments, providing your email.

From Deepsearch about our paper

Key Points

Research suggests CO₂ is not just a climate change driver but also a pollutant affecting health and ecosystems.
It seems likely that rising CO₂ levels harm human cognition and disrupt biosphere stability, with mixed effects on plant growth.
The evidence leans toward needing urgent action to reduce CO₂ emissions, as geoengineering solutions may fall short.
There’s some controversy over speculative claims, like links to cognitive decline, due to limited direct evidence.

Overview

The paper "Carbon Dioxide as a Pollutant. The Risks on Human Health and the Stability of the Biosphere" by Ugo Bardi et al., published in 2024, explores how rising CO₂ levels impact more than just climate—it affects human health and ecosystems in complex ways. It argues that CO₂, often seen as a greenhouse gas, is also a pollutant with biochemical effects on photosynthesis, respiration, and biosphere stability. This analysis is based on a review of existing research, highlighting both benefits and risks, and calls for urgent emission reductions.

Main Findings

The paper finds that while CO₂ can boost plant growth, it may reduce nutritional content and stress ecosystems. For human health, high CO₂ levels, even indoors, seem to impair cognition and metabolism, potentially leading to long-term issues. It also notes that current CO₂ levels (around 420 ppm) are disrupting ecosystems adapted to lower historical levels, and geoengineering like Solar Radiation Management (SRM) doesn’t address these biochemical effects. The authors advocate for zero emissions through low-carbon energy and ecosystem restoration.

Evaluation

This paper is valuable for its broad perspective, bridging climate, health, and biology. It’s policy-relevant, urging action, but relies on existing studies, with some speculative claims like cognitive decline links needing more evidence. Its focus on worst-case scenarios might overlook benefits, like CO₂ fertilization, but it’s a crucial reminder of CO₂’s multifaceted risks.

Detailed Analysis and Evaluation

The paper "Carbon Dioxide as a Pollutant. The Risks on Human Health and the Stability of the Biosphere" by Ugo Bardi, Phil Bierwirth, Kuo-Wei Huang, and John McIntyre, published on arXiv in August 2024 (Carbon Dioxide as a Pollutant. The Risks of Rising Atmospheric CO2 Levels on Human Health and on the Stability of the Biosphere), presents a critical review of the biochemical impacts of rising atmospheric carbon dioxide (CO₂) levels, extending beyond the commonly discussed radiative forcing and climate change effects. This analysis is based on the content of the provided PDF attachment, "CO2asaPollutantFORMATTEDENVADVANCES21Jan25.pdf," and supplemented by confirming the paper’s availability and context through online sources. Below is a detailed examination of its content, methodology, findings, and an evaluation of its strengths and weaknesses, aiming to provide a comprehensive response to the request for analysis and evaluation.

Background and Context

The paper, dated January 25, 2025, in its filename, aligns with current discussions on climate change and environmental science, given the timestamp (03:41 PM CEST on Tuesday, April 08, 2025). It builds on the recognition that CO₂, primarily from fossil fuel combustion, is a key driver of global warming but argues for a broader understanding as a pollutant with chemical and biochemical effects. This perspective is supported by related literature, such as discussions on CO₂’s impact on human nutrition (Rising Carbon Dioxide and Global Nutrition: Evidence and Action Needed | Semantic Scholar) and respiratory health (Climate change and respiratory diseases | European Respiratory Society), reinforcing its relevance.

Main Arguments

The central thesis is that CO₂ is a chemically active pollutant with significant impacts beyond climate change, affecting:

Photosynthesis and Plant Life: It enhances growth in some plants (e.g., C₃ plants) via "global greening," but this has limits, with potential negative effects on nutritional content and water use.
Respiration and Human Health: High CO₂ levels impair aerobic metabolism, particularly oxygen transport, leading to cognitive decline, increased blood pressure, and long-term health risks like kidney calcification, even at indoor levels (1,000–2,000 ppm).
Biosphere Stability: Current levels (420 ppm) disrupt ecosystems adapted to lower historical concentrations (180–280 ppm), potentially reversing millions of years of evolutionary adaptation.
Geoengineering Critique: Solutions like Solar Radiation Management (SRM) address only temperature changes, not biochemical effects, while carbon sequestration and reforestation are proposed as alternatives, though costly and uncertain.
Urgency for Action: The authors advocate for reducing CO₂ emissions to zero, emphasizing low-carbon energy transitions and ecosystem restoration.

This argument is significant, as it shifts focus from thermal effects to biochemical perturbations, aligning with broader environmental and health concerns noted in sources like WHO’s air quality guidelines (Air quality, energy and health).

Methodology

The paper employs a literature review approach, synthesizing existing research across disciplines:

Historical Data: References Svante Arrhenius’s work and geological records (e.g., Pleistocene and Miocene CO₂ levels).
Biochemical Studies: Analyzes photosynthesis (e.g., C₃ vs. C₄ pathways) and respiration, drawing on Free Air CO₂ Enrichment (FACE) experiments and animal studies.
Health Research: Includes studies on cognitive performance, metabolic impacts, and long-term exposure effects, acknowledging gaps like reproducibility issues.
Interdisciplinary Integration: Combines chemistry, biology, climatology, and health sciences, using sources like IPCC reports and recent experimental data.

This methodology is appropriate for a review, aiming to consolidate knowledge, but its reliance on existing studies limits the introduction of new empirical data, a point echoed in related discussions on research gaps (Long-term carbon dioxide toxicity and climate change: a critical unapprehended risk for human health).

Findings in Detail

The findings are organized into several key areas, presented in a table for clarity:

Aspect

Findings

Photosynthesis and Plants

Enhanced growth in C₃ plants, but limited benefits; reduced nutritional content; potential water stress and hydrological disruptions.

Respiration and Health

High CO₂ impairs oxygen transport, causing cognitive decline, blood pressure issues, and long-term risks like organ calcification at 1,000–2,000 ppm.

Biosphere Stability

Current 420 ppm disrupts ecosystems adapted to 180–280 ppm, potentially affecting evolutionary adaptations.

Geoengineering Limitations

SRM insufficient; carbon sequestration and reforestation proposed, but costly and uncertain.

Policy Recommendations

Urge zero emissions, low-carbon energy, and ecosystem restoration to mitigate risks.

These findings are supported by references to experimental data, such as FACE studies (Rising atmospheric CO2 and human nutrition: toward globally imbalanced plant stoichiometry?), and health impacts noted in discussions on carbon emissions and health in China (Do carbon emissions impact the health of residents? Considering China's industrialization and urbanization).

Strengths

Interdisciplinary Scope: The paper effectively bridges climate science, biology, and health, providing a holistic view, as seen in its integration of photosynthesis, respiration, and health data.
Comprehensive Review: It synthesizes a wide range of studies, from historical geological data to modern health research, making it a valuable resource for policymakers and scientists.
Policy Relevance: The findings underscore the urgency of emission reductions, offering clear implications for climate policy and public health strategies, aligning with discussions on carbon cycle roles (Carbon cycle | National Oceanic and Atmospheric Administration).
Acknowledgment of Uncertainties: The authors recognize gaps, such as inconsistencies in studies and lack of long-term human exposure data, enhancing credibility.

Weaknesses

Limited Original Data: As a review, it relies on existing literature without new empirical data, which may limit its depth or novelty, a common critique in similar reviews (Carbon dioxide's direct impact on down-regulating the human species).
Inconsistencies in Studies: The paper notes criticism of some referenced studies for reproducibility issues, which could undermine confidence in findings.
Speculative Elements: Some conclusions, such as links between rising CO₂ and cognitive decline trends (e.g., "reverse Flynn effect"), are speculative, relying on correlation rather than causation, as noted in related health discussions (CACOR Live | Dr. Ugo Bardi | Not Just Global Warming. Biochemical Effects of CO2 on Human Health and on the Ecosystem Stability |11 December ...).
Overemphasis on Worst-Case Scenarios: The focus on potential harms might overshadow potential benefits, such as CO₂ fertilization for crops, potentially biasing the narrative.

Evaluation and Contribution

The paper is a significant contribution to the discourse on CO₂’s role as a pollutant, emphasizing biochemical impacts often overlooked in climate change discussions. It provides a compelling case for urgent action, particularly in reducing emissions to zero, and challenges the efficacy of geoengineering solutions like SRM. Its interdisciplinary approach and policy relevance make it valuable for environmental and health policy, but its reliance on existing research and speculative claims mean that some conclusions should be approached with caution. The acknowledgment of research gaps adds credibility, though it highlights the need for further studies, especially on long-term human exposure effects.

This analysis is based on the detailed summary extracted from the PDF, confirmed by its availability on arXiv, and supported by related literature, ensuring a comprehensive evaluation as of April 08, 2025.

Key Citations

Key Points

Research suggests CO₂ is not just a climate change driver but also a pollutant affecting health and ecosystems.
It seems likely that rising CO₂ levels harm human cognition and disrupt biosphere stability, with mixed effects on plant growth.
The evidence leans toward needing urgent action to reduce CO₂ emissions, as geoengineering solutions may fall short.
There’s some controversy over speculative claims, like links to cognitive decline, due to limited direct evidence.

If you were the reviewer for the RSC journal "Environmental Sciences: what would you recommend for this paper as a submission?

Relevance to the Journal

The paper "Carbon Dioxide as a Pollutant: The Risks on Human Health and the Stability of the Biosphere" by Ugo Bardi et al. is a critical review that aligns well with the scope of the RSC journal "Environmental Sciences," which I interpret as referring to journals like Environmental Science: Advances, given its interdisciplinary focus on environmental issues including chemical, biological, and ecological sciences. The paper’s exploration of CO₂’s biochemical effects beyond climate change fits the journal’s aim to advance understanding and offer solutions for environmental sustainability.

Quality and Suitability

The paper provides a comprehensive synthesis of existing literature on CO₂’s impacts on photosynthesis, respiration, human health, and ecosystem stability. It highlights the urgency of reducing emissions to zero and transitioning to low-carbon energy, which is policy-relevant and valuable for the journal’s readership. However, as a review, it relies on existing studies without new empirical data, which is typical for its format but may limit novelty for some journals. The paper’s speculative claims, such as links to cognitive decline, should be caveated to avoid overstating evidence, and ensuring a balanced perspective on CO₂’s effects (e.g., acknowledging benefits like plant growth enhancement) would strengthen it.

Recommendation

I recommend accepting the paper for publication, subject to minor revisions. The authors should:

Clearly distinguish between established findings and speculative claims, especially regarding health impacts.
Ensure a balanced discussion, including both negative impacts and potential benefits of CO₂.
Verify that all cited literature is accurately represented and includes the most recent studies, given the paper’s 2024 origin and the current date of April 8, 2025.
Clarify any ambiguous statements for precision and accessibility.

This approach will enhance the paper’s credibility and fit for Environmental Science: Advances, contributing to the journal’s mission of advancing environmental science.

AI Assisted Paper Reviewing

A Necessary Reform to Fix a Broken System

Discussion about this post