Generative AI
& medical Chatbots
The Worries ELIZA sparked are resurfacing in AI psychotherapy
In 1966, Stanford psychiatrist Kenneth M. Colby published one of the first conceptual accounts of a chatbot for psychotherapy. It was the same year in which, at MIT, computer-science pioneer Joseph Weizenbaum released his famous natural-language processing program ELIZA, which many regard as the world’s first chatbot. The two concurrent publications marked the starting points of a fierce controversy about the moral limits of automating psychotherapy. With the advent of generative AI, many of the same questions are now reappearing. Revisiting the early debate can therefore inform discussions on the boundaries of AI-based therapy at a time when it is being considered once again.
the future of AI-based clinical ethics consultation
Hein, A.; Meier, L. J. (2026): Data Science in Clinical Ethics Consultation: Analytics, Chatbots, and Specialized Decision Support Systems. In: Handbook of Digital and Experimental Methods in Bioethics, edited by S. Salloch, K. B. Francis, and B. D. Earp. Berlin: Springer [forthcoming]
[forthcoming in 2026]
data-source problems for patient preference prediction
Automatic Patient Preference Predictors are algorithms that use statistically predictive demographic key characteristics to infer an individual’s treatment preferences in health care. Lately, this initial proposal was expanded into a personalised version: instead of relying on demographic data, large language models are to extract values from a variety of sources generated by individual patients. I sketch two potential problems with the two arguably most transformative types of data source on which such an algorithm would be based: the difficulty of differentiating between general and idiosyncratic statements in online activities, and the tension between incentivising users and the faithfulness of the preferences so elicited. The accuracy of personalised predictor algorithms might be much lower than currently expected.
How to embed ethics into healthcare AI?
Integrating artificial intelligence (AI) into critical domains such as healthcare holds immense promise. Nevertheless, significant challenges must be addressed to avoid harm, promote the well-being of individuals and societies, and ensure ethically sound and socially just technology development. Innovative approaches like Embedded Ethics, which refers to integrating ethics and social science into technology development based on interdisciplinary collaboration, are emerging to address issues of bias, transparency, misrepresentation, and more. This paper aims to develop this approach further to enable future projects to effectively deploy it. Based on the practical experience of using ethics and social science methodology in interdisciplinary AI-related healthcare consortia, this paper presents several methods that have proven helpful for embedding ethical and social science analysis and inquiry. They include (1) stakeholder analyses, (2) literature reviews, (3) ethnographic approaches, (4) peer-to-peer interviews, (5) focus groups, (6) interviews with affected groups and external stakeholders, (7) bias analyses, (8) workshops, and (9) interdisciplinary results dissemination. We believe that applying Embedded Ethics offers a pathway to stimulate reflexivity, proactively anticipate social and ethical concerns, and foster interdisciplinary inquiry into such concerns at every stage of technology development. This approach can help shape responsible, inclusive, and ethically aware technology innovation in healthcare and beyond.
Are large language maps or Fuzzy cognitive maps better at doing medical ethics?
Meier, L. J. (2026): Large Language Models versus Fuzzy Cognitive Maps for Solving Moral Dilemmas. Croatian Journal of Philosophy [forthcoming]
Which is better at doing medical ethics: conversational artificial intelligence bots like ChatGPT or tools based on fuzzy cognitive maps? The article compares the performance of chatbots that rely on large language models to that of our own METHAD algorithm. While both tools approach dilemmas in medical ethics through the lens of Beauchamp and Childress’ mid-level principles, ChatGPT and METHAD differ considerably in the format of their inputs and outputs, in their interpretability, and in the kinds of mistakes that they make. An ideal advisory algorithm would combine their characteristics.
ChatGPT’s Mistakes when confronted with clinical dilemmas
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.