PDF
Citation
Print
INTRODUCTION
There is a growing emphasis on identifying and preventing unethical behavior and detrimental research practices in order to enhance the integrity and openness of scientific studies.1 By enhancing the precision, comprehensiveness, and transparency of research reporting, reporting standards play a critical role in the field of health research.12
Furthermore, the significance of reporting standards goes beyond supporting the meticulous evaluation of study methodology and outcome validity, as they also improve the caliber of evidence synthesis for application in the real world of healthcare. Inadequate reporting hinders the reproducibility of research findings, wastes time and resources, and distorts the available data.3 The comprehensiveness of trial reporting in medical publications is improved by reporting guidelines, such as Consolidated Standards of Reporting Trials (CONSORT), the gold standard for randomized controlled trials (RCTs).3 Although there has been a recent drive for more research transparency, there are still large gaps between stated ideas and actual actions. It has been observed that these gaps can be caused by several variables, including traditions, expectations, and practical considerations when publishing scientific data.1
A reporting guideline’s central checklist outlines the crucial data that must be included in a published article.1 These checklists are based on practices and represent the industry’s duty to ensure the quality and safety of processes and products, especially in high-risk circumstances.4 The World Health Organization surgical safety checklist exemplifies how considerable reductions in post-operative complications and fatality rates have been attained.56
This work contributes to a better understanding of past adherence to reporting standards and the influence of artificial intelligence (AI) technological advances on these processes. This review emphasizes the importance of revised guidelines for ensuring open, reproducible, and robustly conducted current scientific publications. It achieves this by looking into the influence of AI technologies, specifically ChatGPT, on data collection and processing, as well as methodologies for research.
Aims of the paper
1. To provide a historical comprehensive overview of the importance and evolution of research reporting standards.
2. To explore the impact of AI on data collection, analysis, and adherence, as well as what recent changes have been made in response to AI-driven research.
3. To assess the role of commonly used statistical software in affecting reporting standards and guidelines.
4. To provide an overview of the implications of writing software on research reporting and publication ethics.
5. To identify the challenges and considerations of adapting research reporting standards in an ethical manner to ensure reproducibility and replicability.
6. To discuss future directions and provide recommendations for improving research reporting standards.
LITERATURE SEARCH
A thorough review of the literature published in PubMed and Scopus electronic databases was performed to extract published manuscripts until March 30th, 2024, following the recommendations published by Gasparyan et al.7 Only English-language sources were considered, using the search terms: “Reporting Standards,” “Artificial Intelligence,” “Statistics Software,” and “Large Language Models.” The inclusion criteria for selecting articles were articles published up to March 30, 2024, indexed in PubMed and Scopus databases, written in English, and relevant to health research and the impact of AI in this field. We excluded articles that did not fulfill those inclusion criteria. Articles were initially screened based on their title and abstract, and full texts were reviewed to ensure relevance. The search results can be found throughout this review.
EVOLUTION OF RESEARCH REPORTING STANDARDS
The growth of reporting standards in medical research has been characterized by a progressive understanding of the value of well-designed trials and open reporting.8 Throughout the twentieth century, there were concerns regarding the quality and knowledge of research methodologies.9 However, it wasn’t until the late twentieth century and early twenty-first century that a substantial focus was placed on the inadequate reporting in published journal papers.89
Early criticisms (20th century)
Research methods developed in the early 20th century reveal weaknesses in statistical analysis and clinical data. Researchers like Halbert Dunn highlight that medical research has been criticized for lacking methodological rigor and statistical logic.10
Growing concerns (late 20th century)
A lack of statistical planning and evaluation in published articles raised concerns about inadequate technique. Studies examining research report quality found significant percentages of methodological errors.11
Introduction of reporting guidelines
To address these concerns, initiatives were taken to enhance reporting standards. In the 1990s, preliminary recommendations were made about the contents of research publications. Eventually, specific reporting guidelines were created, including RCTs, which are made to accommodate various types of investigations.1213
The CONSORT statement
The first RCT reporting flow diagram and checklist were released in 1996. Their main objectives were to ensure the transparency and completeness of trial reports and correct shortcomings in RCT reporting.14
The influence of CONSORT has greatly affected the evolution of reporting requirements for other study types and RCTs. Also, its methodology, which focused more on reporting than study conduct, was used as a model for other reporting guidelines. Accessible evidence was drawn upon where necessary, and adjustments were made over time to accommodate new evidence and viewpoints.8
The Enhancing the Quality and Transparency Of health Research (EQUATOR) Network
This was founded in 2006 to address the issue of inadequate reporting systematically and globally. Its goal was to unite individuals engaged in research reporting to advance the creation, distribution, and adoption of reporting standards.815
The Global Impact of The EQUATOR Network has enabled the promotion of reporting standards, the provision of tools and training, and the expansion of research into research methodologies. Even though it has improved reporting criteria, there are still issues in ensuring these guidelines are widely adopted and followed.1115
HISTORICAL PERSPECTIVE ON RESEARCH REPORTING STANDARDS
Over the course of a century, there have been various comments of concern over the quality of medical research and for almost as long, there has been much debate about the issues surrounding the quality of research reports.8 The Declaration of Helsinki, first published in 1964, is regarded as a foundational work in human research ethics. It was among the first to expressly address moral issues in human subjects’ medical research.16
Research reporting standards have been continuously developed in the field of research.81718 Moments that demonstrate attempts to increase the transparency and rigor of scientific research indicate the development of research reporting standards.819 Guidelines were established due to early detection of reporting inadequacies.18 These benchmarks include realizing the impact of incomplete reporting on reproducibility, working together to find gaps in reporting procedures, and continuously improving standards to stay updated with changing research techniques.20
Developing reporting rules for particular study designs, determining core outcome sets to improve comparability, and including patient viewpoints in health research reporting are examples of critical moments.212223 These benchmarks highlight the iterative process of improvement propelled by the widespread recognition of the significance of thorough and open reporting in the health sciences.2425 Important organizations have greatly helped the establishment and promotion of standards for research reporting.26
RCTs are fundamental to clinical research to treatments to assess treatments’ safety and effectiveness. RCTs, regarded as the gold standard, objectively assess research outcomes by randomly assigning patients to different therapy.27 One outstanding example of how to improve the standard of clinical trial reporting is CONSORT.3 The CONSORT statement was first provided in 1996 and revised in 2010 to guarantee that RCTs are reported without ambiguity. Consensus on Study Design, Methodology, and Performance CONSORT provides internationally accepted principles for reporting individually randomized parallel comparisons between two groups.28
To assist researchers in publishing cross-sectional, cohort, and case-control studies by fulfilling specific requirements for each type of study, Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) was established in 2004. The criteria for observational studies have been raised as a result.29 By making sure that crucial information on the methodology, conclusions, and discussion is accurately reported in scientific publications, STROBE and CONSORT contribute to the improvement of the thoroughness and clarity of study reporting.30
The Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) declaration acts as a guide for both, and PRISMA has had a significant influence on the reporting of both systematic reviews and meta-analyses31; PRISMA provides authors with a list of 27 critical criteria that they can utilize to help them write thorough and trustworthy summaries of study findings. This invites readers to offer a critical assessment of the work. The collaborative efforts of standards-setting bodies and reporting laws improve the caliber and openness of published medical research.32
These standards-setting bodies have established authoritative sources for researchers seeking guidance on transparent and comprehensive reporting across diverse research domains through international cooperation and interdisciplinary engagement.33
In the contemporary research landscape, myriad reporting guidelines cater to the specific needs of various research methodologies and study designs.34 Researchers can access a wealth of structured frameworks to transparently report methods, results, and interpretations.35 The Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) for nonrandomized studies.36 and the diversity and specificity of available reporting guidelines are illustrated by the comprehensive collection by the EQUATOR Network.2
THE EMERGENCE OF AI
AI can be defined in many ways. Some consider it the technology developed to let computers and other machinery operate intelligently.38 In contrast, AI can be viewed as a machine that replaces human labor and produces faster and more effective results for humankind.39 For others, it is perceived as “a system” that can accurately understand outside information, learn from it, and apply that learning to accomplish objectives and activities via adaptable modification.4041
The impact of AI on health research can be highlighted in the following points:
1. AI is a powerful referencing tool that tracks recent literature, finds relevant papers, maps knowledge areas, and generates syntheses on specific research questions. Platforms like Google Scholar, Iris.ai, and Elicit.org enable researchers to navigate literature efficiently, saving time and effort in literature review processes.42
2. AI can improve research designs by assisting researchers in selecting appropriate research methods, validating results, and designing research instruments.42
3. AI can help researchers improve survey research by assisting in various aspects, including survey design, sampling, data cleaning, analysis, and reporting, thus improving the quality and efficiency of the research process.43
The increasing usage of writing software and statistics research has become essential to modern empirical investigations in healthcare sciences.44 Statistical software development has completely changed how researchers and scientists analyze data.44 Because modern software tools are so accurate and efficient, performing extensive statistical analyses by hand is now outmoded and prone to errors.45 Moreover, a clear benefit comes from simplified data interpretation using visual aids like graphs and charts to improve understanding and results dissemination,46 accepting many kinds of data, such as category and numeric values. By incorporating a variety of statistical tests into their frameworks, these tools streamline the computation of results.46 Research has revealed which statistical software is most used and preferred in studies. Because of its extensive capability and easy-to-use interface, the Statistical Package for Social Sciences (SPSS) is often used for epidemiological studies.47 On the other hand, because of its sophisticated commands, Statistical Analysis Systems (SAS) may be difficult for inexperienced users, although offering more flexibility and sophisticated graphical features.48 Another well-known tool is Stata, which provides a strong programming language for data analysis that supports both conventional and unconventional approaches.49
RESEARCH REPORTING STANDARDS IN THE AGE OF AI
AI has rapidly evolved into a widely recognized buzzword, joining discussions across diverse industries and academic disciplines. The utilization of AI machines, exemplified by technologies like ChatGPT, has attracted significant attention in recent years.
Academic writing notably benefits from adopting large language models (LLMs), which are models that produce languages that appear to have been developed by people using computational AI approaches.5051 By assisting writers with writing, grammar, and language usage, they can raise the quality of manuscripts, which results in well-written articles that follow precise guidelines for coherence and clarity and are logically organized.52 LLMs also play a part in breaking down language barriers by offering non-English-speaking writers with a free tool for language editing services.53 and LLMs helps scholars understand challenging topics, review long materials, and produce well-organized summaries.54 Accordingly, literature synthesis is accelerated, enabling a thorough comprehension of complex concepts.
Additionally, LLMs are crucial in streamlining the reporting of research findings by facilitating statistical analysis and guaranteeing accuracy and clarity in presenting results, and they help writers adhere to stringent reporting requirements.5255 Additionally, after a publication, LLMs improve the discoverability of scientific material.52 They also assist writers with the literature evaluation phase,56 and because they can manage enormous volumes of data, they are suitable for applications that require processing huge datasets. 4
Publishers have responded to the launch of LLMS; for example, Science, a prestigious publication, like many others, updated its editorial guidelines to expressly forbid ChatGPT content and make it clear that the software cannot be credited as an author. Editors have raised concerns about researchers relying on ChatGPT and other AI programs while preparing manuscripts.57 They point out likely errors and the risk that literature reviews and findings summaries won’t provide enough context or examination.58 Even with all the hype surrounding ChatGPT, they emphasize how important it is to maintain high standards in academic writing.596061
Despite being educated on enormous datasets and based on transformer architecture, LLMs still have issues, such as producing false information. Multiple research studies conducted in 2023 showed that LLMs could occasionally provide a more accurate rheumatological diagnosis than rheumatologists.62 However, the study also emphasized the dangers of false positives and misinterpreted data.60616263
The Committee on Publication Ethics (COPE) joined the Journal of the American Medical Association (JAMA) and the World Association of Medical Editors (WAME) organizations to conclude that authors utilizing AI tools in manuscript writing, image/graphic creation, or data collection/analysis must transparently disclose the AI tool’s usage and specify the tool in the Materials and Methods section. Authors bear full responsibility for their manuscripts, including AI-generated content, and are accountable for publication ethics violations.6465
THE ROLE OF STATISTICS SOFTWARE IN RESEARCH REPORTING
When designing a clinical trial or laboratory experiment, statistics play a crucial role in determining the design and sample size that will maximize the likelihood of finding effects of clinical or scientific interest. Additionally, statistics are utilized in data analysis to validate conclusions for a larger population.66 Many statistical software, such as SPSS Statistics, SAS, and Microsoft Excel, are currently being used. In an era marked by a surge in healthcare data, statistical analysis is a key player in medical research, offering a robust framework for unraveling complex datasets and extracting meaningful insights.67
Statistics are necessary for published findings to be considered as reputable and for medical research to offer insightful conclusions.68 Researcher examination of data educated decision-making, and creation of visually appealing reports are all made possible by statistical software.69 Statistics allows researchers to see beyond raw data to identify patterns, relationships, and trends, which helps them come to conclusions backed by evidence.70
Using these techniques, real impacts are separated from chance, guaranteeing that the published results are accurate and not just insignificant.71 Statistical software makes group comparisons easier and provides information on risk variables, treatment effectiveness, and other relevant comparisons.72
These software tools are more than just computational tools; they are facilitators that let researchers examine, evaluate, and present data in previously unheard-of ways.7374 The smooth incorporation of these technologies into the research workflow streamlines processes, promoting more accurate and efficient analyses.75
Research outputs’ quality and transparency are supported by the cooperation of statistical software and reporting guidelines.76 By automating sophisticated computations and testing, these technologies reduce the possibility of human error—a critical concern when working with big datasets.77
Statistical software generates standardized output that ensures reporting uniformity while adhering to established norms for reproducibility and openness.78 To enable result replication and bolster the general credibility of the research, researchers should enhance transparency by providing documentation of their software and analysis techniques.79
According to their study design and data type, statistical software tools direct researchers toward the most relevant statistical procedures, encouraging adherence to reporting requirements such as CONSORT and STROBE.80
Guideline updates emphasize increased clarity and transparency in presenting statistical methods, including software versions and specific analysis steps.81 They advocate for clearly reporting uncertainty measures and confidence intervals alongside P values, promoting a comprehensive understanding of research findings.82 There is also a growing emphasis on effective data visualization to enhance clarity and explicit reporting of potential limitations and sources of bias in statistical analysis.83
WRITING SOFTWARE AND ITS IMPACT ON RESEARCH REPORTING
Academic writers must have a fundamental understanding of how to use writing software. Several resources are accessible to help with different areas of writing thanks to the most recent sophisticated breakthroughs in this industry. These resources can benefit non-native English writers, especially regarding tasks like paraphrasing and improving grammar.84 How writing software is developing makes writing easier and makes academic writing much clearer and of higher quality overall. This is in line with the growing use of technology in academic research and communication. For non-Anglophone authors, LLMs such as ChatGPT provides a free solution that has historically prevented them from being on the same playing field as their native speakers. Because many professional language editing services have serious drawbacks, this is especially helpful for researchers from economically disadvantaged areas.85 The latest developments in writing software, particularly those driven by AI, might affect the students. While these technologies are practical and effective, there is reason to worry that the excessive reliance of the next generation on them may compromise their ability to think critically and write intricately.
Is ChatGPT a useful editing tool for people who don't speak English well? Lingard86 created a common method for using ChatGPT as a language editor and critically analyzing its output; they also offered numerous perspectives on applying LLMs as editing instruments. The study concluded that ChatGPT’s impressive text generation capacity is not matched by its editing capacity and advised the use of alternative AI tools, such as Quillbot, Grammarly, and ProWritingAid, which are intended exclusively for language editing and may provide more helpful grammar fixes.
Plagiarism in scholarly works is a serious issue of rising concern with the increased use of writing software. Plagiarism is against the core of academic integrity since it fails to credit the sources of one's ideas as well as the contributions of others.8788 Writers need to reproduce the meaning of the original ideas using their own words and sentence structure, ensuring a clear distinction between their contributions and those of others.8789 There are different terminologies used when it comes to plagiarism. Paraphragiarism occurs when a writer rephrases or paraphrases ideas from a source without giving proper credit or attribution to the original author.88 It may provide the false impression that the writer is the author of the ideas when taken from another source.88 There are many types of plagiarism which can be summarized as90:
1. Plagiarism of ideas: A writer makes another person’s idea, thought, or invention appear as their own without proper acknowledgment; this can be hard to check but is considered a serious offense.
2. Plagiarism of text (direct plagiarism): Copying sections of text from another source without providing citation or using quotation marks, a practice facilitated by technological advancements leading to increased instances of cut-copy-paste practices, is referred to as plagiarism. Plagiarism can occur even when copying small phrases without acknowledgment, and its detection relies on reviewers making careful comparisons.
3. Mosaic plagiarism: When we borrow ideas and opinions from a source, along with using some exact words or phrases, without giving credit to the original author, it can lead to a presentation that feels tangled and mixed up. Properly referencing and citing sources helps to clarify which words are ours and which are borrowed from other writers. This type of plagiarism produces a “confused plagiarized mass” by hiding the differences between the concepts of the source and the original work. However, if we don't do this correctly, it can confuse our readers and raise ethical concerns.89
4. Self-plagiarism: When we reuse large parts of our previous work without properly acknowledging it or citing it, people wonder if we're being transparent and following the rules of publication ethics. Asking permission to reuse our work and giving clear citations can help address these concerns. But suppose we end up with too much overlap or duplication in our articles; in that case, it wastes resources and raises ethical problems, harming the credibility of scholarly publications.
Detecting acts of using AI-driven writing software to generate content that is not original poses a complex problem. Although plagiarism detection tools are available, they might not be precise in identifying content produced by LLMs because of their continuous evolution.9192 Authors should acknowledge using AI tools or writing software, especially if they significantly contributed to the content.
It is suggested by the International Committee of Medical Journal Editors (ICMJE) that journals make it a requirement for authors to indicate whether they made use of AI-assisted tools such as LLMs or chatbots in generating their submitted articles, indicating how AI was utilized in the methods and acknowledgments section while emphasizing human responsibility for accuracy and integrity of AI content.9394 Moreover, all materials should be attributed correctly, authorship credit should not be given to AI, and there should be a declaration of AI-generated text and images on non-plagiarism.93 Failure to do so can be considered a breach of academic integrity and may undermine the credibility of research.
The question of whether employing ChatGPT in academic writing is plagiarism or not depends on how people interpret and view it. There has been a continuing debate within the academic community about using ChatGPT and other AI language models in research papers.95 According to the WAME, plagiarism refers to “…the use of others’ published and unpublished ideas or words (or other intellectual property) without attribution or permission, and presenting them as new and original rather than derived from an existing source.”96
CHALLENGES AND CONSIDERATIONS IN ADAPTING REPORTING STANDARDS IN THE ERA OF AI
Lund and Wang97 found that various ethical considerations in utilizing ChatGPT for academic writing have been identified, encompassing five key areas: bias concerns, privacy risks, autonomy and informed consent, transparency and accountability, and adherence to intellectual property regulations.97 Any language model might incorporate bias, typically resulting from the training process. Ray54 found 28 potential biases that users of ChatGPT may encounter. These biases include, among others:
1. Availability bias: The propensity of ChatGPT to prefer data that is easier to remember or more accessible within its training set has been noted.
2. Groupthink bias: By producing content that mirrors the consensus viewpoints or ideas revealed in its training data. This may reduce the range of views and make it more difficult to investigate opposing or divergent opinions.
3. Commercial bias: Because ChatGPT’s training data primarily comes from the internet, it may represent the objectives and interests of commercial enterprises, which raises the possibility of commercial bias. This may cause the model to produce material that unintentionally promotes goods, services, or brands—even when the user does not mean it.
LLMs such as ChatGPT have notable limitations since they rely on training data and are unable to access current information from the internet, resulting in stagnant knowledge.98 Using ChatGPT to perform the same queries simultaneously produced different conversations, demonstrating the unpredictable nature of AI model behavior.99 The data are scattered, so it isn’t easy to work with or understand. Furthermore, these models cannot communicate with other systems, such as databases or APIs, which restrict their use in real-time applications that need to manipulate data.54 Furthermore, it’s unclear which particular data drives the model’s answers, making it difficult to grasp the basis of the outputs the lack of transparency is worrying.100 These restrictions make it difficult to guarantee accuracy, adaptability, and transparency when using LLMs.
The legal ramifications of ChatGPT are still being investigated.98 Legally speaking, since ChatGPT is not a person,101 who can be held accountable? In medicine, using electronic health record-based data research is a common practice. Patients’ sensitive medical information may be stored and made available to unknown parties by using this data and voluntarily entering it into AI analysis tools.102 Such data must be legally protected before attempting ChatGPT trials. When employing ChatGPT, security precautions such as encryption, access control, secure data storage, and adherence to privacy laws must be implemented to protect patient information.102
It is commonly recognized that LLMs can generate inaccurate and erroneous data.54100103 To ensure reproducibility, the use of LLMs to create content is expected in the current era, so tools that can detect AI-generated or human-generated texts could help ensure that a study is reproducible. However, sophisticated prompts and advanced AI technologies will likely evade these detection techniques. Instead of launching a pointless arms race between AI chatbots and AI chatbot detectors, publishers and the research community should collaborate to find ethical, open, and honest ways to use LLMs.56
Recommendations
Collaboration with specialists from numerous disciplines is essential for comprehending non-scientist viewpoints and foreseeing ethical, legal, and societal ramifications. Social sciences, law, and ethics are a few of these disciplines, but they are not the only ones. Organizations, publications, and the scholarly community must set forth precise rules and moral principles for the use of writing software.
Healthcare practitioners in non-English speaking countries might benefit from writing programmes and editing systems created at universities to help them write for and publish in peer-reviewed journals.104
The appropriate disclosure practices may be involved, original authorship and AI-assisted content generation may be distinguished, and researchers and students may be educated about the responsible use of these tools. The preservation of critical thinking abilities and ethical writing practices must be balanced with the advantages of efficiency to maintain the integrity of academic research and publications.
Park et al.105 created a double checklist of the PRISMA 2020 guidelines, which included descriptions of several often overlooked components as well as actual solutions for statistical analysis mistakes in published research. Using such checklists may increase reproducibility. Moreover, to address the challenges and considerations in adapting reporting standards in the era of AI, here are some concrete recommendations:
FUTURE DIRECTIONS
The EQUATOR Network offers a wide range of reporting standards, this makes it difficult for researchers to select and utilise the right ones, which is why EQUATOR is working to increase knowledge and motivation for correct usage.106 In October 2019, CONSORT and the EQUATOR expanded the CONSORT trial standards to incorporate trials aided by AI, or CONSORT-AI.107 The CONSORT-AI extension suggests expanding the current CONSORT 2010 statement with fourteen new checklist items (eleven extensions and three elaborations). In addition to the essential factors on the CONSORT 2010 checklist, these items were deemed significant enough to warrant routine reporting in clinical trial reports for AI therapies.108 This expansion is primarily meant for researchers and readers reporting on or evaluating clinical trials. However, it could also be helpful for those developing AI interventions when their systems are still in the early stages of validation.109
Sallam et al.110 developed a checklist that can be used to standardize reporting algorithms for AI-based studies in healthcare based on various methodologies and reports. This checklist can be used to standardize the Design and Reporting of Generative Artificial Intelligence-Based Studies in Healthcare Education and Practice. The checklist is known as “METRICS.” This may be the first beneficial step towards developing a well-recognized methodology to standardize reporting in AI-based healthcare research, a rapidly developing field of study. Additionally, they developed the “CLEAR” tool, a checklist intended to evaluate the caliber of health data provided by AI-based models. With the development of these approaches, a global agreement is required to establish reporting standards for every specific area of study.
The requirement for thorough reporting guidelines for clinical trials employing AI therapies is being addressed by the CONSORT-AI and Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT)-AI Steering Group.107 Despite the Food and Drug Administration having approved over thirty AI algorithms, the evidence supporting their efficacy in improving patient outcomes is largely limited to the diagnostic accuracy of these algorithms. The CONSORT and SPIRIT declarations, which disclose study protocols and randomized trials, have greatly improved transparency.111
We believe that the combination of powerful instruments can accelerate breakthroughs at a never-before-seen rate in a few life sciences disciplines, which is why we support their revolutionary potential. However, we also emphasize the importance of rigorous ethical scrutiny to ensure their use aligns with the greater good. To guide the world towards a responsible and balanced reliance on new technologies, journals, and associations should prioritize a full investigation of these ethical problems for the sake of humanity.
Scholarly publication, therefore, needs authors to evaluate and adhere to publication ethics, especially with the development of Open Access journals. These publications undermined evidence-based research by producing articles of such poor scientific quality. However, despite efforts to address the issue, such as adding unethical publications to the list of prohibited journals, posting the statements, and improving indexing criteria, no one solution is effective.112
It is important to note that the findings of this review may have some limitations. Our search was limited to English-written manuscripts, which could introduce language bias and exclude potentially important findings from non-English articles. Additionally, due to the large number of papers found, we were only able to search the PubMed and Scopus databases, potentially overlooking significant findings from other databases.
CONCLUSION
The creation of research reporting standards reflects measures to improve scientific transparency, rigor, and reproducibility. There are still concerns about plagiarism and ethical use even if ChatGPT and other AI applications have benefits for improving writing quality and translated text. The essential nature of collaboration, promptness, and adherence to publication ethics should be considered. Focused and trustworthy insights for research conclusions are ensured by statistical software. Transparency and replicability are enhanced by its collaboration with reporting standards, thus strengthening the basis for medical investigation. Future directions should encompass AI-aided trials as well as standardized reportage approaches. Global researchers must come to a consensus on actions to advance AI-based healthcare research. In short, new issues must be addressed, and technology must be used sensibly and openly by all parties. Together, we can manage challenging circumstances and ensure the public trusts our scientific research.
XML Download