The stakes are considerable. Scoping reviews allow researchers to identify key concepts, theoretical frameworks and knowledge gaps by systematically charting the available literature, and they often determine whether a full systematic review is worth undertaking. Yet the sheer volume of published science makes comprehensive coverage increasingly difficult, particularly in interdisciplinary fields like forest ecology where relevant evidence is scattered across agronomy, ecology and silviculture. The Spanish team set out to determine whether a carefully adapted large language model could shoulder part of that burden without sacrificing the rigor that scientific synthesis demands—a question they frame as urgent, since using generative AI in literature reviews has been categorized as a high-risk practice for research integrity.

Their approach rests on a technical architecture known as Retrieval-Augmented Generation, or RAG. Rather than letting a general-purpose model such as GPT-4o answer from its pretraining alone, the researchers constrained the model to draw exclusively on a curated knowledge base of 33 peer-reviewed papers, stored in a SharePoint library and retrieved through semantic search. The RAG process unfolds in three phases: retrieval, in which the system identifies the most relevant text fragments for a given query; augmentation, in which those fragments are injected into the model’s prompt as supplementary context; and generation, in which the model composes its answer using only the retrieved material. This design dramatically reduces the risk of hallucinated citations and invented findings, because every claim is anchored to a verified source document.

Building the knowledge base was itself a deliberately human enterprise. The team searched Scopus and Web of Science using a structured query that required abstracts to mention trees or forests, to include the terms mixed, mixture or mixing, and to contain the word overyielding, while excluding tropical forests and restricting results to English-language articles published between 2020 and 2024. After removing duplicates and screening out an off-topic genetics paper, 33 documents remained. For each one, the researchers manually extracted the functional trait analyzed, the biome and location, the methods, the ecosystem services addressed and the observed outcome, creating a structured map of the literature that temperate European forests dominate, with Asian studies concentrated in Chinese and Japanese plantations and American work spanning Canada, Colombia and the United States.

The AI agent itself was built with a low-code approach using Microsoft Copilot Studio, a choice the authors emphasize for its accessibility. Constructing a RAG system from scratch demands data scientists, engineers and significant funding—resources rarely available in forest science labs. The low-code route required minimal customization and can be operated by the researchers themselves. The agent’s behavior was defined through a meta-prompt following the PERFECT framework, which specifies the prompt’s Purpose, Elements, Role, Format, Examples, Conditions and Timeframe. This meta-prompt established the agent’s role as a scientific reviewer, its professional tone, its obligation to stay within the curated corpus, and its required output structure based on the PICOT(S) scheme borrowed from clinical trial design, which characterizes Population, Intervention, Comparison, Outcome, Time and Study type.

A second, subtler layer of the design involved grounding. Beyond the RAG knowledge base, the researchers supplied a separate library of three authoritative reference works on mixed-species forests. These grounding documents are never cited in the agent’s responses, but they are available to the model in every iteration, allowing it to acquire the vocabulary, processes and conceptual relationships of the domain—to, in effect, think like an expert. The authors compare this to traditional fine-tuning, but at lower cost and with greater control, and they argue it improves the agent’s ability to disambiguate, prioritize and synthesize information in line with the project’s technical context.

To test the system, the team posed twelve instructions to the agent, covering six research questions about how mixing species with different functional traits affects ecosystem service provision. Half the prompts used a zero-shot approach—concise questions with no added detail—while half used refined-shot prompts tailored with specifics about the species identities in the analyzed forests. The results were revealing. Zero-shot prompts achieved high precision, with a mean success rate of 0.89, meaning the agent usually selected the correct articles, but suffered from low sensitivity, with a mean of 0.36, missing many relevant papers. Refined-shot prompts reduced omissions across all objectives but tended to increase the failure rate, sometimes pulling in related but non-germane articles.

The most striking finding concerned the economics question. For the topic of economical return, the refined prompt produced a complete retrieval failure, with a success rate and sensitivity of zero, apparently because the wording conflated biomass production with monetary terms and generated a query that matched nothing in the corpus. This demonstrates a crucial technical insight about RAG systems: prompt refinement affects not only how information is summarized but which documents are retrieved in the first place. A more specific prompt can fail to surface relevant papers that use different terminology. The authors conclude that prompt quality matters, but that the meta-prompt and retrieval configuration may matter more, and that each research question requires its own careful evaluation of prompt wording.

Validation came from a three-way comparison. The agent’s outputs were checked against the authors’ expert assessment and independently against the commercial model Claude, which was queried about each cited article without access to the agent’s knowledge base or grounding documents. All three evaluations agreed completely, supporting the team’s first hypothesis that a RAG-based language model can produce outputs valid for scientific scoping reviews. The main limitation, they found, was article retrieval rather than the correctness of retrieved information—the agent knew what to say about the papers it found but sometimes failed to find all the papers it should have.

The broader implications extend well beyond forestry. The methodology is explicitly transferable to other scientific domains, and it arrives as commercial tools such as Scopus AI and LeapSpace begin to mediate researchers’ access to literature—tools that, unlike the customizable agent described here, offer limited opportunities to inspect or modify the retrieval layer and meta-prompt. The authors call for interdisciplinary collaboration among computer scientists, Earth domain scientists and AI specialists to build accessible, domain-specific RAG systems at reasonable cost. They also acknowledge open questions: the environmental footprint of generative AI, the variability of model outputs across sessions, the exclusion of foundational older papers imposed by the curated time window, and the need for systematic experiments on how different levels of human expertise affect reliability. What the study demonstrates, however, is a practical middle path—neither full automation nor wholesale rejection—in which machines do the heavy lifting of retrieval and drafting while human scientists retain control over corpus selection, system configuration and final judgment. For a discipline that manages ecosystems on century timescales, that partnership may prove the most consequential mixing effect of all.

Subject of Research: Human–AI hybrid methodology using retrieval-augmented large language models for scoping reviews of ecosystem services in mixed forests

Article Title: Integrating human and artificial intelligence for conducting scoping reviews on ecosystem services in mixed forests

Article References: Bravo, F., Ruano, I., Martín, D., & Ordóñez, C. (2026). Integrating human and artificial intelligence for conducting scoping reviews on ecosystem services in mixed forests. Discover Artificial Intelligence, 6(1), Article 1438. https://doi.org/10.1007/s44163-026-02464-w

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02464-w

Keywords: generative artificial intelligence, large language models, retrieval-augmented generation, scoping review, mixed forests, plant functional traits, ecosystem services, overyielding, forest ecology, prompt engineering, research methodology, GPT-4o

The team designed an elegant experiment around five varieties of pepper, Capsicum annuum: two wild ancestors and three modern elite cultivars. Each variety was grown either alone in monoculture or intercropped with maize, Zea mays, one of the world’s most dominant cereal crops. Crucially, the researchers ran the whole experiment twice over, once under normal moisture conditions and once under drought, creating a full factorial design that allowed them to separate the effects of domestication history, cropping system, and water availability. This mattered because the outcome of any plant interaction is not fixed; the stress-gradient hypothesis in ecology holds that neighbors that harm each other under benign conditions can switch to helping each other when the environment turns harsh, as shade from one plant conserves soil moisture for another.

The results confirmed that moisture dramatically rewires the pepper-maize relationship. Under normal watering, intercropping actually suppressed pepper growth, with maize presumably winning the contest for light and resources. Under drought, however, the picture flipped: pepper responses became predominantly neutral or even facilitative, and every wild and modern pepper-maize combination achieved overyielding, meaning the relative yield total exceeded one. In practical terms, the mixed stands produced more combined biomass per unit area than the two species grown separately would have. For a world facing growing water scarcity, especially in major agricultural regions of China where this work was conducted, the finding that drought can transform a competitive interaction into a productive partnership is genuinely striking.

Yet the study’s most provocative insight concerns domestication. Wild pepper varieties showed higher log response ratios, a standard metric of how a plant’s performance changes between monoculture and mixture, and intercrops involving wild varieties achieved higher relative yield totals than those with modern cultivars. In other words, the wild plants were better at turning a neighbor’s presence into an advantage. Interestingly, this relative superiority of wild peppers did not vary significantly with moisture, suggesting that the domestication penalty is baked into the plants themselves rather than being an artifact of any particular watering regime. Modern breeding, optimized for high yields in monoculture fields with ample inputs, may have inadvertently selected against traits that make crops good teammates.

To understand the mechanism, the researchers measured seventeen functional traits spanning leaves and roots, the biological machinery that determines how a plant acquires light, water, and nutrients. Under normal moisture, modern pepper varieties displayed a distinctive combination in monoculture: acquisitive leaf traits, the fast-growth leaf economics typical of bred crops, paired with conservative root traits. More tellingly, modern varieties showed greater intercropping-induced variability in root traits than wild ancestors when water was plentiful. The team quantified this plasticity using multidimensional hypervolumes, a modern ecological technique that treats a plant’s suite of traits as a cloud of points in multi-dimensional space and measures how that cloud expands or shifts when conditions change.

Drought told a different story. When water became scarce, intercropped modern peppers showed reduced variation in root chemical and morphological traits compared with their well-watered counterparts. This contraction of the root trait space suggests that modern cultivars lose part of their capacity to reconfigure their underground foraging strategy precisely when reconfiguration matters most. Roots are the hidden half of intercropping success; complementary root depths, architectures, and nutrient-acquisition strategies are among the main drivers of overyielding in cereal-legume and cereal-vegetable mixtures worldwide. If domestication has narrowed the plasticity of root-acquisition traits, modern peppers may simply be less able to negotiate a division of labor with maize below ground.

The study fits into a growing body of evidence that crop domestication disrupts biodiversity effects. Previous work has shown that domestication erodes beneficial interactions in barley-faba bean mixtures, shifts competitive ability in durum wheat, and alters root trait syndromes across many crop types. What makes the new pepper-maize study distinctive is its explicit focus on how moisture modulates these domestication effects through the lens of functional trait plasticity. Plasticity, the ability of a single genotype to express different phenotypes in different environments, is a double-edged trait: it is advantageous when environments vary predictably and costly when they do not. The authors argue that domestication was associated with trait-specific changes in moisture-dependent plasticity, particularly in root-acquisition traits, and that this may constrain how modern varieties respond to the simultaneous pressures of competition and facilitation.

The implications for breeding are tantalizing. If the wild ancestors of pepper carry alleles that promote beneficial root plasticity and facilitative interactions, those traits could theoretically be reintroduced into elite germplasm, either through conventional crossing with wild relatives or through genomic selection. The authors explicitly suggest that reintroducing beneficial wild traits may improve crop performance in intercropping systems under variable environments. This is part of a broader conversation in agroecology about back-crossing to the wild, a strategy that has been proposed for restoring lost genetic diversity in everything from disease resistance to drought tolerance. The pepper study adds a new dimension to that debate: what is being lost may not only be stress tolerance itself, but the behavioral flexibility that allows crops to cooperate with their neighbors.

There are, of course, caveats. The experiment involved a single partner crop, maize, and a handful of pepper varieties, and biomass responses in a controlled setting do not always translate directly to grain or fruit yield in farmers’ fields. The relative advantages of wild varieties also did not shift with moisture in the statistical tests, hinting that the domestication effect is robust but perhaps not moisture-specific in the way the trait data alone might suggest. Still, the combination of trait measurements, hypervolume analysis, and yield metrics makes this one of the most mechanistically complete examinations of domestication’s cost to intercropping to date, and it arrives at a moment when intercropping is being promoted as a pillar of sustainable intensification from smallholder farms in Asia and Africa to experimental systems in Europe and North America.

For now, the message for growers and breeders alike is nuanced but clear. Intercropping pepper with maize remains a promising strategy, especially under water-limited conditions where every combination tested achieved overyielding. But the full potential of these ancient partnerships may be locked behind traits that modern breeding left behind in the wild. As climate change makes rainfall less predictable and water scarcity more common, the plants best suited to the fields of the future may be those that have not forgotten how to share. Reconnecting modern crops with their wild heritage, the study suggests, could be the key to unlocking it.

Subject of Research: Effects of crop domestication on functional trait plasticity and intercropping performance of pepper with maize under varying moisture conditions

Article Title: Impaired intercropping effect of domesticated pepper with maize under variable moisture conditions: a functional trait perspective

Article References: Ye, J., Liu, Y., Yi, W., Zhao, L., Liu, Y., Guo, L., Tang, J., Hu, L., & Chen, X. (2026). Impaired intercropping effect of domesticated pepper with maize under variable moisture conditions: a functional trait perspective. Plant and Soil. https://doi.org/10.1007/s11104-026-09086-2

Image Credits: AI Generated

DOI: 10.1007/s11104-026-09086-2

Keywords: crop domestication, intercropping, pepper, maize, drought stress, functional traits, phenotypic plasticity, root traits, overyielding, relative yield total, agroecology, plant breeding