Search

In this article, we explore the potential of enhancing academic prose and idea generation by fine-tuning a large language model (here, GPT-3) on one's own previously published writings: AUTOGEN ("AI Unique Tailored Output GENerator"). We develop, test, and describe three distinct AUTOGEN models trained on the prior scholarly output of three of the current authors (SBM, BDE, JS), with a fourth model trained on the combined works of all three. Our AUTOGEN models demonstrate greater variance in quality than the base GPT-3 model, with many outputs outperforming the base model in format, style, overall quality, and novel idea generation. As proof of principle, we present and discuss examples of AUTOGEN-written sections of existing and hypothetical research papers. We further discuss ethical opportunities, concerns, and open questions associated with personalized academic prose and idea generators. Ethical opportunities for personalized LLMs such as AUTOGEN include increased productivity, preservation of writing styles and cultural traditions, and aiding consensus building. However, ethical concerns arise due to the potential for personalized LLMs to reduce output diversity, violate privacy and intellectual property rights, and facilitate plagiarism or fraud. The use of coauthored or multiple-source trained models further complicates issues surrounding ownership and attribution. Open questions concern a potential credit-blame asymmetry for LLM outputs, the legitimacy of licensing agreements in authorship ascription, and the ethical implications of coauthorship attribution for data contributors. Ensuring the output is sufficiently distinct from the source material is crucial to maintaining ethical standards in academic writing. These opportunities, risks, and open issues highlight the intricate ethical landscape surrounding the use of personalized LLMs in academia. We also discuss open technical questions concerning the integration of AUTOGEN-style personalized LLMs with other LLMs, such as GPT-4, for iterative refinement and improvement of generated text. In conclusion, we argue that AUTOGEN-style personalized LLMs offer significant potential benefits in terms of both prose generation and, to a lesser extent, idea generation. If associated ethical issues are appropriately addressed, AUTOGEN alone or in combination with other LLMs can be seen as a potent form of academic enhancement. [ABSTRACT FROM AUTHOR]

The authors discuss the article "AUTOGEN: A Personalized Large Language Model for Academic Enhancement--Ethics and Proof of Principle" by S. Porsdam Mann and colleagues. Topics include issues on the application of AI in education, claim by the researchers on AUTOGEN, and ways that the widespread and indiscriminate use of AUTOGEN may harm the integrity of scholarly writing.

In this paper, we contend with whether we still need traditional ethics education as part of healthcare professional training given the abilities of chatGPT (generative pre-trained transformer) and other large language models (LLM). We reflect on common programmatic goals to assess the current strengths and limitations of LLMs in helping to build ethics competencies among future clinicians. Through an actual case analysis, we highlight areas in which chatGPT and other LLMs are conducive to common bioethics education goals. We also comment on where such technologies remain an imperfect substitute for human-led ethics teaching and learning. Finally, we conclude that the relative strengths of chatGPT warrant its consideration as a teaching and learning tool in ethics education in ways that account for current limitations and build in flexibility as the technology evolves. [ABSTRACT FROM AUTHOR]

The application of artificial intelligence and machine learning (ML) technologies in healthcare have immense potential to improve the care of patients. While there are some emerging practices surrounding responsible ML as well as regulatory frameworks, the traditional role of research ethics oversight has been relatively unexplored regarding its relevance for clinical ML. In this paper, we provide a comprehensive research ethics framework that can apply to the systematic inquiry of ML research across its development cycle. The pathway consists of three stages: (1) exploratory, hypothesis-generating data access; (2) silent period evaluation; (3) prospective clinical evaluation. We connect each stage to its literature and ethical justification and suggest adaptations to traditional paradigms to suit ML while maintaining ethical rigor and the protection of individuals. This pathway can accommodate a multitude of research designs from observational to controlled trials, and the stages can apply individually to a variety of ML applications. [ABSTRACT FROM AUTHOR]

Conversational artificial intelligence (CAI) presents many opportunities in the psychotherapeutic landscape—such as therapeutic support for people with mental health problems and without access to care. The adoption of CAI poses many risks that need in-depth ethical scrutiny. The objective of this paper is to complement current research on the ethics of AI for mental health by proposing a holistic, ethical, and epistemic analysis of CAI adoption. First, we focus on the question of whether CAI is rather a tool or an agent. This question serves as a framework for the subsequent ethical analysis of CAI focusing on topics of (self-) knowledge, (self-)understanding, and relationships. Second, we propose further conceptual and ethical analysis regarding human-AI interaction and argue that CAI cannot be considered as an equal partner in a conversation as is the case with a human therapist. Instead, CAI's role in a conversation should be restricted to specific functions. [ABSTRACT FROM AUTHOR]

The concept of personhood has been central to bioethics debates about abortion, the treatment of patients in a vegetative or minimally conscious states, as well as patients with advanced dementia. More recently, the concept has been employed to think about new questions related to human-brain organoids, artificial intelligence, uploaded minds, human-animal chimeras, and human embryos, to name a few. A common move has been to ask what these entities have in common with persons (in the normative sense), and then draw conclusions about what we do (or do not) owe them. This paper argues that at best the concept of "personhood" is unhelpful to much of bioethics today and at worst it is harmful and pernicious. I suggest that we (bioethicists) stop using the concept of personhood and instead ask normative questions more directly (e.g., how ought we to treat this being and why?) and use other philosophical concepts (e.g., interests, sentience, recognition respect) to help us answer them. It is time for bioethics to end talk about personhood. [ABSTRACT FROM AUTHOR]

The author comments on the article "AUTOGEN: A Personalized Large Language Model for Academic Enhancement--Ethics and Proof of Principle" by S. Porsdam Mann ad colleagues which showed a technique for fine-tuning the large language model (LLM) GPT-3. Topics include considerations for academic research introduced by LLMs, concerns raised by the authors around consent in the context of fine-tuned LLMs, and challenges over potential plagiarism and authorship misrepresentation.

The article examines whether one should understand a fine-tuned large language model (LLM) trained on one's own past writings as a form of academic enhancement of oneself. Topics discussed include reason that current LLMs are sometimes called hallucinating bullshitters, the concept AUTOGEN-LLMs, and argument of David Chalmers, one of the two main characters behind the extended mind thesis, on LLMs.

The author offers a response to the remarks by S. Porsdam Mann et al on the possible pitfalls from the increasing use of generative artificial intelligence (GenAI) for academic enhancement combined with the publish or perish imperative prevalent in academia. Topics include the risk that perverse incentives might arise from such a combination, the challenge of keeping up with published literature, and the concern that GenAI might have a homogenizing effect on the writing style of users.

The authors propose four general guidelines for using artificial intelligence (AI) in the bioethics classroom such as transparency, antidiscrimination, privacy, and accountability. They stress the need of professors and instructors to accept responsibility and outcomes for AI generated products. They also offer a suggestion to prepare the next generation of bioethicists.

The authors comment on an article by V. Rahimzadeh and colleagues on the role of large language models (LLMs) in ethics education. Topics discussed include how LLMs can support the learning process, how LLMs can best facilitate ethics education, and analysis of learning students how to use LLMs in the manner of Rahimzadeh and colleagues' case study.

The authors comment on the study "Algorithms for Ethical Decision Making in the Clinic," wherein L. Meier and colleagues describe their approach to developing machine intelligence to support ethical decision-making in the clinic. Topics include poorly defined problem statement and inadequate consideration of constraints, heterogeneity and representatives of data sources used for training, and programmability of the ethical framework.

The authors comment on the proof-of-concept study by L. Meier and colleagues, published within the issue wherein they built an algorithm to aid technical decision making. Topics include the failure of the study authors to convincingly describe the problem that they are trying to solve, their failure to explain why an artificial intelligence (AI) driven system is the appropriate tool to aid ethical decision-making, and the approach of the project to problem-solving.

This open peer commentary refers to the article "Toward a Framework for Assessing Privacy Risks in Multi-Omic Research and Databases," by Charles Dupras and Eline M. Bunnik, that was published within the issue. Topics discussed include artificial intelligence (AI) methodologies for analyzing biomedical data, human-like behavior of machines, and ethical concerns from the power of AI when applied to human tissue repositories.

The article considers large language models (LLMs) like OpenAI's ChatGPT as a novel application of machine learning (ML) techniques. Topics discussed include risk that are expected to be exacerbated by misuse of ML in bioscience, ideas about policy responses aimed at mitigating the foregoing risks, and suggestion on future public policy research.

The author comments on an article by I. G. Cohen which showed that ChatGPT has been focused on potential risks while artificial intelligence (AI)-based chatbots could also positively empower patients. Topics include other potential benefits to ChatGPT that deserve the attention of bioethecists, how bias is a major issue in the development of AI in medicine and beyond, and suggestion to put more focus on images when studying tools like ChatGPT.

The article discusses the ethical implications of artificial intelligence (AI) platforms enabled by large language models (LLMs) for medicine. Topics include expectations on chatbots in the area of mental health, bioethical questions raised by chatbots, and evolving socio-technological practices that raise new bioethical questions of AI-human interactions in psychotherapy.

In the last several months, several major disciplines have started their initial reckoning with what ChatGPT and other Large Language Models (LLMs) mean for them – law, medicine, business among other professions. With a heavy dose of humility, given how fast the technology is moving and how uncertain its social implications are, this article attempts to give some early tentative thoughts on what ChatGPT might mean for bioethics. I will first argue that many bioethics issues raised by ChatGPT are similar to those raised by current medical AI – built into devices, decision support tools, data analytics, etc. These include issues of data ownership, consent for data use, data representativeness and bias, and privacy. I describe how these familiar issues appear somewhat differently in the ChatGPT context, but much of the existing bioethical thinking on these issues provides a strong starting point. There are, however, a few "new-ish" issues I highlight – by new-ish I mean issues that while perhaps not truly new seem much more important for it than other forms of medical AI. These include issues about informed consent and the right to know we are dealing with an AI, the problem of medical deepfakes, the risk of oligopoly and inequitable access related to foundational models, environmental effects, and on the positive side opportunities for the democratization of knowledge and empowering patients. I also discuss how races towards dominance (between large companies and between the U.S. and geopolitical rivals like China) risk sidelining ethics. [ABSTRACT FROM AUTHOR]

The authors comment on the article "Epistemic rights and responsibilities of digital simulacra for biomedicine" by M. Cho and N. Martinez-Martin which outlines challenges and limitations of incorporating digital simulacra in healthcare. They discuss the nonhypothesis-driven approach and the principle of data minimalization, the bias-reducing potential of the approach, and the nonhypothesis-driven approach against the traditional biomedical scientific methods and the logic of clinical reasoning.

Big data and AI have enabled digital simulation for prediction of future health states or behaviors of specific individuals, populations or humans in general. "Digital simulacra" use multimodal datasets to develop computational models that are virtual representations of people or groups, generating predictions of how systems evolve and react to interventions over time. These include digital twins and virtual patients for in silico clinical trials, both of which seek to transform research and health care by speeding innovation and bridging the epistemic gap between population-based research findings and their application to the individual. Nevertheless, digital simulacra mark a major milestone on a trajectory to embrace the epistemic culture of data science and a potential abandonment of medical epistemological concepts of causality and representation. In doing so, "data first" approaches potentially shift moral attention from actual patients and principles, such as equity, to simulated patients and patient data. [ABSTRACT FROM AUTHOR]

The article focuses on framework for the ethical evaluation of conversational artificial intelligence (CAI) in psychotherapy. Topics discussed include finding the right balance between the status of a tool and an agent, understanding technologies as social phenomena and materialized artifacts created in complex processes.

The author comments on an article by V. Rahimzadeh and colleagues which discussed the virtues and vices of employing ChatGPT in ethics education for health professionals. Topics discussed include a classic dilemma picked by the authors in medical ethics, prima-facie principles relied on by ChatGPT in interpreting the case, and analysis of ChatGPT's handling of dilemmas in medical ethics.

The authors comment on the article "What Should ChatGPT mean for bioethics," by I. Glenn Cohen which explores the bioethical implications of open artificial intelligence's (AI's) chatbot ChatGPT and the use of similar Large Language Models (LLMs). Topics discussed include key elements of a basic interpersonal communication (IPC) model, ChatGPT as a communication tool, and ChatGPT as a bioethics communicator.

The author comments on an essay by S. Porsdam Mann and colleagues about the potential of personalized large language models (LLMs) to create professional-grade academic writing. Topics include claim made by Porsdam Mann and colleagues on LLMs, the opacity of large language models (LLMs), and reason that the use of personalized LLMs like AUTOGEN should be considered a concern.

The authors comment on the article "Epistemic rights and responsibilities of digital simulacra for biomedicine" by M. Cho and N. Martinez-Martin which highlights the problems of bias and misrepresentation leading to misdirected care, epistemic injustice and marginalization of certain populations. Topics include forms of concerns on the existential threat posed by technology and requirements in the ethical development and application of digital simulacra and other emerging technologies in care.

The article discusses the potential role of conversational artificial intelligence (CAI) in psychotherapy. Topics discussed include CAI's inability to take normative stance and lack of heterogeneity, framework to CAI's use in psychotherapy and the potential deception involved in forming an empathy-based relationship.

The article discusses the status of conversational artificial intelligence (CAI) in psychotherapy on a continuum between tool and agent. Topics discussed include important ethical implications for patients, physicians, and treatment decisions generated by different conceptions of AI, lacks certain core features of agency in CAI and the impact of agential status of CAI on its ability to offer therapeutic support.

The article discusses about the traditional features of and requirements for a therapeutic relationship between a therapist and a patient and the failure of conversational artificial intelligence (CAI) in that regard. Topics discussed include the normative requirements governing the human-CAI collaboration, importance of CAI systems being assigned agency and a model of shared agency between CAI systems and humans.

Machine intelligence already helps medical staff with a number of tasks. Ethical decision-making, however, has not been handed over to computers. In this proof-of-concept study, we show how an algorithm based on Beauchamp and Childress' prima-facie principles could be employed to advise on a range of moral dilemma situations that occur in medical institutions. We explain why we chose fuzzy cognitive maps to set up the advisory system and how we utilized machine learning to train it. We report on the difficult task of operationalizing the principles of beneficence, non-maleficence and patient autonomy, and describe how we selected suitable input parameters that we extracted from a training dataset of clinical cases. The first performance results are promising, but an algorithmic approach to ethics also comes with several weaknesses and limitations. Should one really entrust the sensitive domain of clinical ethics to machine intelligence? [ABSTRACT FROM AUTHOR]

The article presents the discussion on commentators taking the time for commenting on the work. Topics include data sovereignty and collaboration in the research process respecting indigeous forms of knowledge and independence; and facilitating the development of research being sensitive to the harms and benefits of AI/ML research.

The authors comment on an article by Lukas J. Meir and colleagues which offered the promise of a pathway for resolving clinical bioethical problems using an artificial intelligence (AI) interface. They discuss the issues that arise in evaluating whether the AI program, METHAD accomplishes the intended initial goal and possible role of fuzzy cognitive maps (FCMs) such as the ones used in METHAD in the enacting of ethics consultations.

The authors comment on an article by M. Pyrrho, L. Cambraia, and V. F. de Vasconcelos on protecting privacy while optimizing the use of data. They talk about the increase in the possibilities for collecting, storing, and processing of data, the key principles in European data protection legislation, and the warning by Pyrrho, Cambraia, and de Vasconcelos on the conversion of data into economic value.

The article discusses the rise of the bioethics artificial intelligence (AI). Topics include the release of Delphi, an artificial AI trained to make general moral decisions released by the Allen Institute for Artificial Intelligence in October 2021; the four prima facie Principles of Biomedical Ethics, and comparison between AI systems Delphi and METHAD.

The article presents the discussion on machine learning in health care posing new challenges for appropriating evaluation for safe use in clinical care. Topics include introduction of algorithmic biases particularly in relation to race, gender, and class; and identifying a number of potential ethical concerns inherent in medical research employing AI.