EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Human and planetary health implications of AI-generated diet plans: a comparative assessment using the EAT-Lancet framework

Başak Öney, Solmaz Ece Yılmaz

Frontiers in Sustainable Food Systems · 09.11.2026 · 10

DOI: 10.3389/fsufs.2026.1958219 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Frontiers in Sustainable Food Systems, online erschienen am 09.11.2026. Verfasst von Başak Öney und 1 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Artificial intelligence (AI) is increasingly being used to support dietary planning, with growing interest in generating recommendations that are both nutritionally adequate and environmentally sustainable. This study compared meal plans generated by three AI platforms, OpenAI ChatGPT (GPT 5.5 Instant), Google Gemini (Gemini 3 Flash), and Anthropic Claude (Claude Opus 4.8), across different clinical scenarios using the energy adjusted EAT–Lancet Diet Index 2.0 (ELI 2.0) together with carbon and water footprint indicators.

Methods

Diet plans were generated by ChatGPT, Gemini, and Claude for five clinical scenarios (healthy, obesity, type 2 diabetes, hypertension, and hypercholesterolemia) using an identical standardized prompt. Each platform generated three independent meal plans for each scenario. Nutrient composition was analyzed using BEBİS software, diet quality was assessed using the ELI 2.0, and environmental sustainability was evaluated by calculating carbon and water footprints. Differences between platforms and clinical scenarios, as well as output reproducibility, were analyzed.

Results

Energy adjusted ELI 2.0 scores were comparable across the three AI platforms ( p = 0.240). Water footprint differed significantly between platforms ( p = 0.004), whereas carbon footprint differences were not significant after multiple-comparison correction ( p = 0.093). Significant differences were also observed in several nutrient outcomes, and repeated meal plan generation demonstrated limited within platform reproducibility.

Discussion

Comparable adherence to the Planetary Health Diet did not necessarily indicate comparable environmental sustainability across AI generated meal plans. Future evaluations should incorporate both nutritional and environmental outcomes.

Zitieren

APA

Öney, B., & Yılmaz, S. E. (2026). Human and planetary health implications of AI-generated diet plans: a comparative assessment using the EAT-Lancet framework. Frontiers in Sustainable Food Systems, 10. https://doi.org/10.3389/fsufs.2026.1958219

Vancouver

Öney B, Yılmaz SE. Human and planetary health implications of AI-generated diet plans: a comparative assessment using the EAT-Lancet framework. Frontiers in Sustainable Food Systems. 2026;10. doi:10.3389/fsufs.2026.1958219

BibTeX
@article{ney2026,
  author = {Öney, Başak and Yılmaz, Solmaz Ece},
  title = {Human and planetary health implications of AI-generated diet plans: a comparative assessment using the EAT-Lancet framework},
  journal = {Frontiers in Sustainable Food Systems},
  year = {2026},
  volume = {10},
  doi = {10.3389/fsufs.2026.1958219}
}
Kostenlos starten →