Academic Integrity Review: "Stochastic Biomechanical Modeling of Human-Powered Electricity Generation" (Energies, 2025)

Verdict

🔴 Confirmed (high confidence): Multiple independent indicators of AI-generated or fabricated content, including verbatim text duplication, internally inconsistent core metrics, physiologically implausible efficiency values, and questionable reference formatting. Final determination requires institutional ethics review.

Key findings

• Verbatim text duplication (Section 4.6, pp. 15-16): Two consecutive paragraphs are nearly identical, repeating the same cohort fitness range (0.88–0.95) and recommendation (75–150 N force).
• RMSE inconsistency: Reported as 3.52 W in Section 4.1 text versus 8.3 ± 1.7 W in Table 3 for the same stochastic model.
• Maximum power contradiction: Maximum Power Archetype listed as 175.5 W in Table 1 but described as 427 ± 29 W per operator in the Section 4.5 text.
• Arithmetic error in data points: Section 3.1 states 89 trials × 7 participants = 623 trials; Section 4.1 claims 267 × 7 = 1869 individual measurement points, introducing an unexplained 267 multiplier.
• Participant age conflict: Section 3.1 states ages 20–31; Sections 5.2 and 5.6 state 22–35.
• Implausible efficiency: Reported biomechanical and metabolic efficiency of η = 0.891 ± 0.016 contradicts established physiology (human muscle metabolic efficiency peaks near ~25%).
• Suspiciously fast timeline: Received 3 August 2025; Accepted 9 September 2025 (~1 month), despite claimed 623 human-subject trials with lactate, EMG, and cardiopulmonary measurements.
• Reference formatting anomalies (unverified): Several entries lack authors and/or titles (e.g., Refs [1], [8], [33], [36], [58], [59]).

Evidence highlights

• Direct copy-paste, p. 15–16: "While our validation cohort demonstrated fitness levels of 0.88–0.95... For typical sedentary individuals, we recommend reducing the operational envelope to 75–150 N force..." appears twice with negligible variation.
• Table 3 vs. text: RMSE = 3.52 W (Section 4.1) vs. 8.3 ± 1.7 W (Table 3, "This Work (Stochastic)").
• Power Archetype: 175.5 W (Table 1) vs. 427 ± 29 W (Section 4.5 narrative).
• Trial accounting: 89 × 7 = 623 (Section 3.1) vs. 267 × 7 = 1869 (Section 4.1).
• Age range: 20–31 (Section 3.1) vs. 22–35 (Sections 5.2, 5.6).
• Efficiency claim: ηbio = ηmetabolic = 0.891 ± 0.016 (Table 2), inconsistent with physiological literature.
• Reference anomalies: "1. Nat. Rev. Methods Primers 2023, 3, 39." and similar entries missing authors/titles (⚠ not independently verified from the supplied PDF).

Notes

• DOI: 10.3390/en18184821 — preserved as published.
• The corresponding author's primary research focus (deep-sea submersibles) contrasts sharply with the human-physiology methodology described, which may suggest honorary authorship or ghost writing; this cannot be confirmed from the manuscript alone.
• The 1-month submission-to-acceptance window is unusual for a study claiming 623 human-subject physiological trials and warrants editorial scrutiny.
• Reference-list issues were flagged by the original reviewer but could not be directly cross-checked against the supplied PDF excerpt; treat as pending verification.
• All findings should be formally submitted via PubPeer and to the journal editorial office; the authors' institution should also be notified to verify IRB/ethics approval and raw data availability.
• This AI-assisted assessment is not a definitive finding of misconduct; final adjudication requires institutional investigation.