When Intuition Meets the Algorithm: Medico-Legal Implications of Artificial Intelligence-Driven Decision-Making in Orthopedics

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for revising your research. In its current form, the research is more understandable and fit for purpose. It is suitable for publication as it stands.

Best

Author Response

In the manuscript, revisions are made with Word track changes. Text additions are written in different colors and underlined, while deletions are shown with strikethrough. For improving readability, new sentences are also highlighted in yellow, as are the new references in the References section. A clear manuscript is provided in pdf format.

Single Reviewers’ answers  

Reviewer 1  

Dear Authors, 

 Thank you for revising your research. In its current form, the research is more understandable and fit for purpose. It is suitable for publication as it stands. 

 Best.  

Many thanks for your current and previous support in the text improvement. 

Best Regards 

 

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript explores the application of artificial intelligence in orthopedic practice and discusses its medico-legal implications. While the topic is timely and relevant, the manuscript presents substantial methodological and scientific limitations that significantly weaken its contribution.

• The study is essentially a narrative review without a clearly defined methodological framework. There is no description of the literature search strategy, databases consulted, time frame, inclusion or exclusion criteria, or study selection process. The absence of transparency in literature selection prevents reproducibility and makes it impossible to assess potential selection bias.
• The manuscript lacks any quantitative synthesis or comparative analysis. No numerical data, performance metrics, pooled outcomes, or structured comparisons between AI-assisted and conventional orthopedic decision-making are provided. Although sensitivity, specificity, and predictive performance are mentioned in a general manner, these concepts are not supported by summarized evidence, tables, or figures.
• The paper does not provide a novel scientific contribution. The manuscript largely reiterates existing discussions found in prior narrative reviews and policy-oriented papers.
• The technical aspects of artificial intelligence are treated superficially. Key elements such as algorithmic architectures, training strategies, validation methodologies, external testing, bias mitigation, and clinical deployment pipelines are not addressed in sufficient depth. The manuscript does not distinguish between different AI model types or explain their respective strengths, limitations, and applicability in orthopedic settings.
• The manuscript suffers from structural repetition and a diffuse focus. Similar arguments regarding algorithmic opacity, professional responsibility, and medico-legal uncertainty are repeated across multiple sections without substantive development.
• Finally, although the medico-legal discussion is extensive, it remains largely normative and speculative, lacking empirical grounding. The proposed recommendations and future perspectives are not supported by outcome-based evidence or validated frameworks, further limiting the manuscript’s scientific impact.

Author Response

Reviewer 2  

The manuscript explores the application of artificial intelligence in orthopedic practice and discusses its medico-legal implications. While the topic is timely and relevant, the manuscript presents substantial methodological and scientific limitations that significantly weaken its contribution. 

 The study is essentially a narrative review without a clearly defined methodological framework. There is no description of the literature search strategy, databases consulted, time frame, inclusion or exclusion criteria, or study selection process. The absence of transparency in literature selection prevents reproducibility and makes it impossible to assess potential selection bias. 

We included a brief methodological paragraph to clarify the overall approach of the work; however, we specify that this contribution is not intended to represent a systematic review nor a comprehensive review of artificial intelligence applications in orthopedics. Indeed, there was neither a systematic intent in literature selection nor an aim to achieve thematic saturation. The selection of articles was carried out by the first author based on a qualitative assessment of contributions on the application of AI in orthopedics – prioritizing more recent one - that, according to his scientific experience and sensitivity, were better suited to highlight specific medico-legal implications, as well as to support a broader reflection on overall medico-legal issues, including professional and organizational liability. 

In light of these characteristics, it may be more appropriate to frame the present contribution as an Opinion paper, aimed at stimulating discussion and offering an interpretative perspective on emerging issues, rather than as a review in the strict sense. 

     The manuscript lacks any quantitative synthesis or comparative analysis. No numerical data, performance metrics, pooled outcomes, or structured comparisons between AI-assisted and conventional orthopedic decision-making are provided. Although sensitivity, specificity, and predictive performance are mentioned in a general manner, these concepts are not supported by summarized evidence, tables, or figures. 

    The paper does not provide a novel scientific contribution. The manuscript largely reiterates existing discussions found in prior narrative reviews and policy-oriented papers. 

    The technical aspects of artificial intelligence are treated superficially. Key elements such as algorithmic architectures, training strategies, validation methodologies, external testing, bias mitigation, and clinical deployment pipelines are not addressed in sufficient depth. The manuscript does not distinguish between different AI model types or explain their respective strengths, limitations, and applicability in orthopedic settings.  

We reiterate the above: the aim was to provide a medico-legal discussion of some specific applications and, subsequently, of their broader implications, with the purpose of initiating a debate on the evolving approaches to the patient, to disease, and above all to therapeutic decision-making and its underlying assumptions, from which professional responsibility arises. And an evolving approach difficult to reverse or to meaningfully compare with previous methods. 

     The manuscript suffers from structural repetition and a diffuse focus. Similar arguments regarding algorithmic opacity, professional responsibility, and medico-legal uncertainty are repeated across multiple sections without substantive development.  Finally, although the medico-legal discussion is extensive, it remains largely normative and speculative, lacking empirical grounding. 

Thank you very much. We have revised the text and sought to more clearly address, first, specific issues related to the applications of AI in orthopedics and subsequently develop a doctrinal reflection on the overall impact. Clearly, judicial decisions, rulings, or other legal guidance, as well as medico-legal case law, are currently lacking, thus precluding an analysis of real-world practice, the approach, described in the introduction and methods, is exploratory in nature, focusing on the elements underlying responsibility and the identification of causal relationships, necessarily reflecting authors’ perspectives. 

  The proposed recommendations and future perspectives are not supported by outcome-based evidence or validated frameworks, further limiting the manuscript’s scientific impact. 

We agree and have revised the text accordingly. The aim was to highlight, from a personal medico-legal perspective, which areas appear to be critical and how they might be addressed. 

 Many thanks for your support  

Best Regards 

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for the privilege of reviewing the review entitled: +When Intuition Meets the Algorithm: Medico-Legal Implications of Artificial Intelligence-Driven Decision-Making in Orthopedics” submitted to bioengineering-

 

The article describes a review of facts related to the implementation of LLMs and MLLMs in the field of medicine. The article covers a wide range of medico-legal aspects connected to the implementation of AI driven tools in orthopedic surgery.

Authors describe possible medico-legal implications of the use of AI and how the use of these tools may impact surgical decision making and shift responsibilities in this process from doctors to other health care employees and institutions.

 

Unfortunately, I miss a chapter regarding strengths and limitations of this review.

One bias is obvious to me, as authors state that users do not understand the functions of AI driven tools. However, they use this tool for improvement of the text of this review. The text is crucial to express their thoughts and here authors allow ChatGPT5 to modify their text focusing on the medico-legal aspects of using for example ChatGPT5 in orthopedics.

 

I am disappointed about the missing new facts regarding the use of AI in orthopedics.

The review describes different medico-legal aspects of the implementation of AI driven tools. Given the poor experiences of lawyers investigating the hidden training methods of AI-algorithms and their unsuccessful intellectual property claims against big tech companies I hypothesize that any legal action regarding the responsibility of AI algorithms in medico-legal law cases will be unsuccessful. As you mention some nations now are going to implement first legislations on the use of AI algorithms in medicine. Are there any legal experiences when using theses new laws in the EU?

 

I also miss a chapter focusing on legal ethical issues regarding the implementation of AI driven algorithms in orthopedics.

 

I think authors should focus in depth on the real impact of AI on orthopedics nowadays.

What are real achievements obtained so far by the implementation of AI tools in orthopedics? Is it only about writing the hospital records, is it language recognition, is it establishing the diagnosis by evaluation of patient data? What has been implemented successfully so far?

 

Another big issue is data safety. Many health data of patients are examined by AI-algorithms, and the servers are located elsewhere and data safety is a big concern. Where do orthopedic surgeons store the data of their patients? Are data of orthopedic patients saved in a cloud outside their hospital or even abroad? Are their data used to train AI algorithms for free?

 

Another issue are costs of LLM or MLLM implementation for patients and physicians. Who is going to pay for these technologies?

 

You ask for training of physicians before implementation of the AI tools. At which costs and time expenses? Frankly, have you had a training before you applied ChatGPT5 for the first time?

 

You advocate for a collaboration of universities with big tech companies to improve AI tools and recommend prospective multi center studies to investigate the medico-legal implications of AI algorithms applied to support orthopedic surgeons in their work. Your recommendation would require that both institutions are interested in such collaborations. Are you sure this will work?

 

Some parts of the manuscript are written wordy and superficially. I think description of real-world examples for every aspect could improve the message to the reader. I think you should be able to find examples from Italian courts where AI algorithms were examined in medico-legal cases.

The messages of some parts of the manuscript sound simple and not informative at all. Maybe you ask ChatGPT5 to revise the manuscript at least 2 more times to write a more concise review.

Author Response

Reviewer 3  

Thank you for the privilege of reviewing the review entitled: “When Intuition Meets the Algorithm: Medico-Legal Implications of Artificial Intelligence-Driven Decision-Making in Orthopedics” submitted to bioengineering. 

 The article describes a review of facts related to the implementation of LLMs and MLLMs in the field of medicine. The article covers a wide range of medico-legal aspects connected to the implementation of AI driven tools in orthopedic surgery. Authors describe possible medico-legal implications of the use of AI and how the use of these tools may impact surgical decision making and shift responsibilities in this process from doctors to other health care employees and institutions. Unfortunately, I miss a chapter regarding strengths and limitations of this review.  

Thank you very much. We have revised the text and implemented the introduction with methods and limitations. Clearly, judicial decisions, rulings, or other legal guidance, as well as medico-legal case law, are currently lacking, thus precluding an analysis of real-world practice. The approach, described in the introduction and methods, is therefore exploratory in nature, focusing on the elements underlying responsibility and the identification of causal relationships, necessarily reflecting authors’ perspectives. In light of these characteristics, it may be more appropriate to frame the present contribution as an Opinion paper, aimed at stimulating discussion and offering an interpretative perspective on emerging issues, rather than as a review in the strict sense. 

 One bias is obvious to me, as authors state that users do not understand the functions of AI driven tools. However, they use this tool for improvement of the text of this review. The text is crucial to express their thoughts and here authors allow ChatGPT5 to modify their text focusing on the medico-legal aspects of using for example ChatGPT5 in orthopedics.   

With respect, we disagree; in order to understand innovations, it is necessary to use them. The primary aim of this work, however, was to discuss the distinction between use for language review and use for therapeutic decision-making, and the resulting liability implications that need in-depth analysis. 

 I am disappointed about the missing new facts regarding the use of AI in orthopedics.  The review describes different medico-legal aspects of the implementation of AI driven tools. Given the poor experiences of lawyers investigating the hidden training methods of AI-algorithms and their unsuccessful intellectual property claims against big tech companies I hypothesize that any legal action regarding the responsibility of AI algorithms in medico-legal law cases will be unsuccessful. As you mention some nations now are going to implement first legislations on the use of AI algorithms in medicine. Are there any legal experiences when using theses new laws in the EU? 

We obviously nowadays lack decisions, rulings, or other legal guidance, as well as medico-legal case law, thus precluding an analysis of real-world practice, and we agree with your considerations. The aim was to globally address those aspects, not immediate for clinicians and healthcare managers.  

 I also miss a chapter focusing on legal ethical issues regarding the implementation of AI driven algorithms in orthopedics. 

 I think authors should focus in depth on the real impact of AI on orthopedics nowadays. 

 What are real achievements obtained so far by the implementation of AI tools in orthopedics? Is it only about writing the hospital records, is it language recognition, is it establishing the diagnosis by evaluation of patient data? What has been implemented successfully so far? 

 Another big issue is data safety. Many health data of patients are examined by AI-algorithms, and the servers are located elsewhere and data safety is a big concern. Where do orthopedic surgeons store the data of their patients? Are data of orthopedic patients saved in a cloud outside their hospital or even abroad? Are their data used to train AI algorithms for free? 

Another issue are costs of LLM or MLLM implementation for patients and physicians. Who is going to pay for these technologies? 

Thank you very much for the suggestions. The perspective of our manuscript, however, was that of the medico-legal professional and professional liability assessment, which differs from ethical-deontological aspects and from purely legal issues related to data protection and cybersecurity, as these lie outside the perimeter of professional responsibility in healthcare setting and related forensic evaluation. The same applies to the economic aspect, which is associated with any technological innovation, but falls within the realm of health economics and healthcare policy, requiring separate expertise and remaining independent from the scope of medico-legal analysis. 

 You ask for training of physicians before implementation of the AI tools. At which costs and time expenses? Frankly, have you had a training before you applied ChatGPT5 for the first time? 

 We agree; however, before applying AI in the medical record or for generating evidence that forms clinical guidelines, appropriate training programs are essential – and in Italy mandatory by law-, as with any new tool, along with adequate information provided to the patient on these applications. In the event of harm, liability rests with the person employing the tool, just as it would for a newly implanted prosthesis performed incorrectly due to insufficient training in its proper use. 

  You advocate for a collaboration of universities with big tech companies to improve AI tools and recommend prospective multi center studies to investigate the medico-legal implications of AI algorithms applied to support orthopedic surgeons in their work. Your recommendation would require that both institutions are interested in such collaborations. Are you sure this will work? 

Certainly, as some university research centers and major technology companies have already done, as overall device security directly affects financial security and profitability. 

 

Some parts of the manuscript are written wordy and superficially. I think description of real-world examples for every aspect could improve the message to the reader. I think you should be able to find examples from Italian courts where AI algorithms were examined in medico-legal cases. 

 The messages of some parts of the manuscript sound simple and not informative at all. Maybe you ask ChatGPT5 to revise the manuscript at least 2 more times to write a more concise review. 

Thank you very much. We have revised the text and sought to more clearly address. Clearly, judicial decisions, rulings, or other legal guidance, as well as medico-legal case law, are currently lacking, thus precluding an analysis of real-world practice, the approach, described in the introduction and methods, is exploratory in nature. 

Many thanks for your support  

Best Regards 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have adequately addressed my concerns by clearly reframing the manuscript as an exploratory opinion paper rather than a review. While the contribution does not provide empirical or quantitative evidence, its value lies in stimulating medico-legal discussion. I therefore recommend acceptance as an Opinion Paper after minor editorial revisions to ensure conceptual clarity and avoid ambiguity in article type.

Comments on the Quality of English Language

The manuscript is written in generally understandable academic English; however, it would strongly benefit from professional language editing to improve clarity, reduce sentence length, and avoid unnecessary repetition, particularly in the discussion sections.

Author Response

We would like to thank the Editors and Reviewers for the comments and suggestions which gave us the opportunity to make the manuscript more consistent and clearer. While we acknowledge some limitations in the text that we have tried to address through revisions, as well as its value as an Opinion paper, we have done our best to improve the text in the points adequately addressed by the reviewers. 

We have answered the Reviewers’ comments in the specific online section. In the manuscript, revisions are made with Word track changes. Text additions are written in different colors and underlined, while deletions are shown with strikethrough. For improving readability, new sentences are also highlighted in yellow, as are the new references in the References section. A clear manuscript is provided in pdf format. 

Single Reviewers’ answers  

Reviewer 2  

Comments and Suggestions for Authors: The authors have adequately addressed my concerns by clearly reframing the manuscript as an exploratory opinion paper rather than a review. While the contribution does not provide empirical or quantitative evidence, its value lies in stimulating medico-legal discussion. I therefore recommend acceptance as an Opinion Paper after minor editorial revisions to ensure conceptual clarity and avoid ambiguity in article type. 

 

Comments on the Quality of English Language: The manuscript is written in generally understandable academic English; however, it would strongly benefit from professional language editing to improve clarity, reduce sentence length, and avoid unnecessary repetition, particularly in the discussion sections. 

Many thanks. We will further highlight to Editors and Editorial Staff the “opinion paper” format for proper change.  

We have further performed an overall text revision to address language and editing issues. 

Best Regards 

Reviewer 3 Report

Comments and Suggestions for Authors

Thank you for revison of the manuscript and your response. 

Author Response

Thank you for revision of the manuscript and your response. 

Many thanks.