Since their public launch, a little over a year ago, large language models (LLMs) have inspired a flurry of analysis about what their implications might be for medical ethics, and for society more broadly. 1 Much of the recent debate has moved beyond categorical evaluations of the permissibility or impermissibility of LLM use in different general contexts (eg, at work or school), to more fine-grained discussions of the criteria that should govern their appropriate use in specific domains or towards certain (...) ends. 2 With each passing week, it seems more and more inevitable that LLMs will be a pervasive feature of many, if not most, of our lives. It would not be possible—and would not be desirable—to prohibit them across the board. We need to learn how to live with LLMs; to identify and mitigate the risks they pose to us, to our fellow creatures, and the... (shrink)

When making substituted judgments for incapacitated patients, surrogates often struggle to guess what the patient would want if they had capacity. Surrogates may also agonize over having the (sole) responsibility of making such a determination. To address such concerns, a Patient Preference Predictor (PPP) has been proposed that would use an algorithm to infer the treatment preferences of individual patients from population-level data about the known preferences of people with similar demographic characteristics. However, critics have suggested that even if such (...) a PPP were more accurate, on average, than human surrogates in identifying patient preferences, the proposed algorithm would nevertheless fail to respect the patient’s (former) autonomy since it draws on the ‘wrong’ kind of data: namely, data that are not specific to the individual patient and which therefore may not reflect their actual values, or their reasons for having the preferences they do. Taking such criticisms on board, we here propose a new approach: the Personalized Patient Preference Predictor (P4). The P4 is based on recent advances in machine learning, which allow technologies including large language models to be more cheaply and efficiently ‘fine-tuned’ on person-specific data. The P4, unlike the PPP, would be able to infer an individual patient’s preferences from material (e.g., prior treatment decisions) that is in fact specific to them. Thus, we argue, in addition to being potentially more accurate at the individual level than the previously proposed PPP, the predictions of a P4 would also more directly reflect each patient’s own reasons and values. In this article, we review recent discoveries in artificial intelligence research that suggest a P4 is technically feasible, and argue that, if it is developed and appropriately deployed, it should assuage some of the main autonomy-based concerns of critics of the original PPP. We then consider various objections to our proposal and offer some tentative replies. (shrink)

As Cohen (2023) describes, the discourse around ChatGPT has been focused on potential risks while AI-based chatbots could also positively empower patients. There are other potential benefits to Cha...

Cohen’s article (2023) on the significance of ChatGPT for bioethics suggests that little is known about the development of generative AI (“GAI”) in China and other national markets. It warns about...

The potential of using personalized Large Language Models (LLMs) or “generative AI” (GenAI) to enhance productivity in academic research, as highlighted by Porsdam Mann and colleagues (Porsdam Mann...

OpenAI’s Chat Generative Pre-training Transformer (ChatGPT), Google’s Bard and other generative artificial intelligence (GenAI) technologies can greatly enhance the capability of healthcare profess...

In their Target Article, Rahimzadeh et al. (2023) discuss the virtues and vices of employing ChatGPT in ethics education for healthcare professionals. To this end, they confront the chatbot with a moral dilemma and analyse its response. In interpreting the case, ChatGPT relies on Beauchamp and Childress’ four prima-facie principles: beneficence, non-maleficence, respect for patient autonomy, and justice. While the chatbot’s output appears admirable at first sight, it is worth taking a closer look: ChatGPT not only misses the point when (...) applying the principle of non-maleficence; its response also fails, in several places, to honour patient autonomy – a flaw that should be taken seriously if large language models are to be employed in ethics education. I therefore subject ChatGPT’s reply to detailed scrutiny and point out where it went astray. (shrink)

Contrary to how it may seem when we observe its output, an [LLM] is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to...

Artificial intelligence (AI), large language models (LLMs), such as Open AI’s ChatGPT, have a remarkable ability to process and generate human language but have also raised complex and novel ethica...

In the target article, Porsdam Mann and colleagues (2023) broadly survey the ethical opportunities and risks of using general and personalized large language models (LLMs) to generate academic pros...

Generative artificial intelligence (GenAI) promises to revolutionize data-driven fields (Milmo 2023). Building on decades of large language modeling (LLM) (Toner 2023), GenAI can collect, harmonize...

In their recent paper “AUTOGEN: A Personalized Large Language Model for Academic Enhancement—Ethics and Proof of Principle,” Porsdam Mann et al. (2023) demonstrate a technique for fine-tuning the l...

In this reply to our commentators, we respond to ethical concerns raised about the potential use (or misuse) of personalized LLMs for academic idea and prose generation, including questions about c...

Cohen (2023), Rahimzadeh and colleagues (2023), and Porsdam Mann and colleagues (2023) have written thorough and well-canvassed pieces about the ethical and conceptual challenges of large language...

Large Language Models (LLMs) “assign probabilities to sequences of text. When given some initial text, they use these probabilities to generate new text. Large language models are language models u...

Scientific research is supposed to acquire or generate knowledge, but such a purpose would be severely undermined by instances of research misconduct (RM) and questionable research practices (QRP). RM and QRP are often framed in terms of moral transgressions by individuals (bad apples) whose aberrant acts could be made conducive by shortcomings in regulatory measures of organizations or institutions (bad barrels). This notion presupposes, to an extent, that the erring parties know exactly what they are doing is wrong and morally (...) culpable, but had nonetheless proceeded to commit wrongful acts. However, a confession of intent to deceived is often not readily admitted by perpetrators of RM. I posit that beyond the simplistic notion of conscious moral transgression, deficits in epistemic virtues and/or the prevalence of epistemic vices have important roles to play in initiating and driving RM/QRP. For the individual perpetrator, deficits in epistemic virtues could lead to or amplify errors in one’s desperate attempt to be accomplished or to excel, and pushes one across the ethical line or down the slippery slope of misconduct. Likewise, a lack of epistemic virtue within perpetrators’ institution or organization could make it conducive for deceitful acts and suppress indications and warning signs for the former. Furthermore, epistemic vices exhibited by reviewers, editors and journals could also promote RM/QRP. In this view, epistemic failings, rather than widespread moral deficiencies of individuals within the research ecosystem, may underlie the prevalence of RM/QRP. (shrink)

Porsdam Mann and colleagues provide an overview of opportunities and risks associated with the use of personalized large language models (LLMs) for text production in bio)ethics (Porsdam Mann et al...