Machine learning in medicine: Addressing ethical challenges

By Effy Vayena, A. Blasimme, I. G. Cohen. PLoS Medicine, 2018

708 citations recorded by the source corpus.

Effy Vayena and colleagues argue that machine learning in medicine must offer data protection, algorithmic transparency, and accountability to earn the trust of patients and clinicians.

Subfields: General Health Professions, Health Informatics, Public Health, Environmental and Occupational Health

View Paper: https://www.semanticscholar.org/paper/a38b3183a6aef62eed68946a416e31d5aa90da59

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In General Health Professions

In Health Informatics

In Public Health, Environmental and Occupational Health

Citations in the corpus

What Clinicians Want: Contextualizing Explainable Machine Learning for Clinical End Use (2019, background): “, 2018b), and high stakes (Vayena et al., 2018) are some of the technical barriers to practical adoption.”
Artificial intelligence in clinical and genomic diagnostics (2019, methodology): “…by the development of: (i) best practices for the interpretability of predictions to protect patient autonomy and shared decision-making; (ii) fairness standards to minimize disparities induced by machine bias; and (iii) ad hoc guidance to allow for continuous improvement of the algorithms [83].”
Machine learning and deep learning—A review for ecologists (2022, neutral): “…m /t e r m s - a nd - c ond iti on s ) on W il e y O n li n e L i b r a r y f o r r u l e s o f u s e ; OA a r ti c l e s a r e gov e r n e d by t h e a pp li ca b l e C r ea ti v e C o mm on s L i ce n s e will emerge, as they have in other fields (e.g. Hardt et al., 2016 ; Vayena et al., 2018 ).”
Fairness of artificial intelligence in healthcare: review and recommendations (2023, background): “Respecting patient autonomy and protect‑ ing confidential medical information is the foundation of ethical AI implementation, which can only be achieved by addressing important issues related to data privacy [94, 95].”
"Hello AI": Uncovering the Onboarding Needs of Medical Practitioners for Human-AI Collaborative Decision-Making (2019, background): “For example, users may resist adopting a tool if they do not understand its capabilities, its intended use, or its utility over existing practices [48, 72, 74].”
Hurtful words: quantifying biases in clinical contextual word embeddings (2020, background): “–4, 2020, Toronto, ON, Canada Zhang and Lu, et al. on healthcare data may exhibit such biases [12, 13, 69], and it is important to take measures that minimize the impact of bias on predictive systems [48, 65]. In this work, we characterize the biases that can be operationalized in clinical predictio”
Initial validation of the general attitudes towards Artificial Intelligence Scale (2020, background): “Vayena, Blasimme, and Cohen (2018) explored what could be done in response to a majority of the UK public feeling uncomfortable with the use of AI andmachine learning in medical settings.”
Use of AI-based tools for healthcare purposes: a survey study from consumers’ perspectives (2020, background): “Method of data collection for AI may increase risks as AI systems need huge datasets, and patients are concerned that their personal information will be collected without their knowledge [30].”
Artificial Intelligence and Machine Learning in Pathology: The Present Landscape of Supervised Methods (2019, background): ““Quality data” for AI/ML training applications must include accurate, precise, complete, and generalizable information.(6) Laboratory data are often assumed to be sufficiently accurate and precise by both health-care providers and researchers.”
Setting the future of digital and social media marketing research: Perspectives and research propositions (2020, background): “Risks related to security (cyber intrusion and privacy), economic aspects (e.g., job displacement) and performance (e.g., errors and bias, black box and explainability issues) can be minimized by valuing human intelligence and skills (Vayena et al., 2018).”
Fairness without Harm: Decoupled Classifiers with Preference Guarantees (2019, background): “…attaining improved accuracy subject to these preference guarantees represents an important direction for future research in fair machine learning given growing calls for such methods in medicine (see e.g., Ferryman & Pitcan, 2018; Popejoy & Fullerton, 2016; Vayena et al., 2018; Chen et al., 2019).”
Artificial Intelligence and the Implementation Challenge (2019, background): “This point raises the overarching issue of the extent of automation that is possible through applications of ML, linked to speculation about whether ML will mostly augment or actually replace health care providers’ work [1,30].”
A governance model for the application of AI in health care (2019, background): “8 Unrepresentative data can occur due to, for example, societal discrimination (eg, poor access to health care) and relatively small samples (eg, minority groups).”
The ethics of AI in health care: A mapping review. (2020, background): “…(2019) stress, uptake of algorithmic-decision-making tools by the clinical community is highly unlikely until this liability question is resolved (Vayena et al., 2018a, 2018b), which could result in the overarching ethical concern raised in the introduction—that of a significant missed…”