Page 24 - AIH-2-4

Artificial Intelligence in Health AI editorial policy ethics

publication policies and innovative data governance journals protect not only the rigor of science but also the
frameworks that balance scientific transparency with the well-being of vulnerable patients.
ethical imperatives unique to this field. Looking ahead, future research should explore

8. Calls for reform: Elevating the standards interdisciplinary innovations to enhance the robustness,
of peer review interpretability, and clinical utility of AI models in
mental and medical health. Emerging computational
To ensure the safe and effective integration of AI into frameworks 46-49 – such as those based on circular bipolar
clinical practice, scientific publishing – especially in clinical complex intuitionistic fuzzy linguistic information, Frank
journals – must reform its approach to reviewing AI and power aggregation operators, and MABAC models –
ML research. To that end, the following recommendations have demonstrated success in fields, such as renewable
are proposed: energy analysis and wireless communications. In addition,
• Expert reviewers for AI methodologies: Journals should approaches employing neuro-fuzzy, complex propositional
engage data science and AI experts to identify technical picture fuzzy Sugeno–Weber power aggregation and
flaws and verify the reproducibility, transparency, and fractal mathematics, including superior Mandelbrot sets,
robustness of the models. offer promising avenues for managing uncertainty and
• Transparent model evaluation: Manuscripts must improving model transparency. 50-57 While these advanced
provide explicit details regarding model training, techniques have yet to be widely applied in mental or
data handling, and algorithm performance while medical health AI, their adaptation holds potential to
addressing issues, such as class imbalance, bias, and address critical methodological challenges, including class
interpretability. imbalance, model interpretability, and generalizability.
• Encouraging open data and code: To facilitate Integrating such innovations could complement editorial
reproducibility, journals should promote the sharing reforms, pushing the field toward more reliable, ethical,
of data and code, enabling independent verification and clinically impactful AI and ML applications.
and improvement of AI models.
This perspective highlights systemic failures in editorial
• Dedicated spaces for AI methodological critiques: oversight and offers concrete recommendations to reform
Creating sections devoted to methodological peer review processes – reforms essential to maintaining
discussion can encourage healthy academic discourse
and improve the quality of published research. trust in both AI research and its real-world applications.
• Ethical and clinical considerations: All AI-driven Without such change, the promise of AI risks becoming
overshadowed by preventable harm and eroded confidence.
studies should include mandatory sections on ethics— Addressing these challenges is not optional; it is a critical
analyzing informed consent, privacy, and potential responsibility that the scientific community and clinical
harm—to ensure safe and responsible applications in publishers must urgently embrace to protect both patients
clinical settings.
and the integrity of mental health research.
9. Final thoughts: Upholding scientific rigor Acknowledgments
and ethical standards
I extend my heartfelt gratitude to Sean Harty, my first
As AI continues to permeate healthcare, the imperative for mentor and a lifelong friend. I met Sean in 1997, and
rigorous, methodologically sound research grows ever more his immediate recognition of my aptitude for computer
urgent. Inaccurate or insufficiently validated AI models programming and network design changed the trajectory
risk fatal errors—misclassifying suicide risk, withholding of my life. Over the years, I quite literally followed him
necessary care, or prompting harmful interventions. across three different companies—a testament to his
These are not abstract concerns; they are life-or-death exceptional technical skill, leadership, and integrity. Sean
consequences of editorial decisions made today. was not only a guiding force in my early IT and network
Clinical journals serve as critical gatekeepers of scientific engineering career but also a constant sounding board,
integrity, and they must adapt to the challenges posed by always offering his time, wisdom, and a whiteboard to
the complexity and novelty of AI-driven methodologies. help map through complex ideas. His mentorship went far
Only through independent, transparent, and technically beyond professional development; it laid the groundwork
informed peer review can the scientific community ensure for how I think, problem-solve, and lead. Now approaching
that AI tools are deployed ethically, effectively, and safely 30 years of service as Chief Information Officer in local
in clinical settings. By embracing robust methodological state government, Sean remains one of the most influential
critique rather than dismissing it as “overly technical,” figures in my life and a treasured friend of 28 years.

Volume 2 Issue 4 (2025) 18 doi: 10.36922/AIH025210049