Dataset Information

From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices.

ABSTRACT: The debate about the ethical implications of Artificial Intelligence dates from the 1960s (Samuel in Science, 132(3429):741-742, 1960. https://doi.org/10.1126/science.132.3429.741 ; Wiener in Cybernetics: or control and communication in the animal and the machine, MIT Press, New York, 1961). However, in recent years symbolic AI has been complemented and sometimes replaced by (Deep) Neural Networks and Machine Learning (ML) techniques. This has vastly increased its potential utility and impact on society, with the consequence that the ethical debate has gone mainstream. Such a debate has primarily focused on principles-the 'what' of AI ethics (beneficence, non-maleficence, autonomy, justice and explicability)-rather than on practices, the 'how.' Awareness of the potential issues is increasing at a fast rate, but the AI community's ability to take action to mitigate the associated risks is still at its infancy. Our intention in presenting this research is to contribute to closing the gap between principles and practices by constructing a typology that may help practically-minded developers apply ethics at each stage of the Machine Learning development pipeline, and to signal to researchers where further work is needed. The focus is exclusively on Machine Learning, but it is hoped that the results of this research may be easily applicable to other branches of AI. The article outlines the research method for creating this typology, the initial findings, and provides a summary of future research needs.

SUBMITTER: Morley J

PROVIDER: S-EPMC7417387 | biostudies-literature | 2020 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices.

Morley Jessica J Floridi Luciano L Kinsey Libby L Elhalal Anat A

Science and engineering ethics 20191211 4

The debate about the ethical implications of Artificial Intelligence dates from the 1960s (Samuel in Science, 132(3429):741-742, 1960. https://doi.org/10.1126/science.132.3429.741 ; Wiener in Cybernetics: or control and communication in the animal and the machine, MIT Press, New York, 1961). However, in recent years symbolic AI has been complemented and sometimes replaced by (Deep) Neural Networks and Machine Learning (ML) techniques. This has vastly increased its potential utility and impact on ...[more]

PMID: 31828533

Similar Datasets

Project description:BackgroundThere is an increasing number of interactive web-based advance care planning (ACP) support tools, which are web-based aids in any format encouraging reflection, communication, and processing of publicly available information, most of which cannot be found in the peer-reviewed literature.ObjectiveThis study aims to conduct a systematic review of web-based ACP support tools to describe the characteristics, readability, and quality of content and investigate whether and how they are evaluated.MethodsWe systematically searched the web-based gray literature databases OpenGrey, ClinicalTrials.gov, ProQuest, British Library, Grey Literature in the Netherlands, and Health Services Research Projects in Progress, as well as Google and app stores, and consulted experts using the following eligibility criteria: web-based, designed for the general population, accessible to everyone, interactive (encouraging reflection, communication, and processing of information), and in English or Dutch. The quality of content was evaluated using the Quality Evaluation Scoring Tool (score 0-28-a higher score indicates better quality). To synthesize the characteristics of the ACP tools, readability and quality of content, and whether and how they were evaluated, we used 4 data extraction tables.ResultsA total of 30 tools met the eligibility criteria, including 15 (50%) websites, 10 (33%) web-based portals, 3 (10%) apps, and 2 (7%) with a combination of formats. Of the 30 tools, 24 (80%) mentioned a clear aim, including 7 (23%) that supported reflection or communication, 8 (27%) that supported people in making decisions, 7 (23%) that provided support to document decisions, and 2 (7%) that aimed to achieve all these aims. Of the 30 tools, 7 (23%) provided information on the development, all of which were developed in collaboration with health care professionals, and 3 (10%) with end users. Quality scores ranged between 11 and 28, with most of the lower-scoring tools not referring to information sources.ConclusionsA variety of ACP support tools are available on the web, varying in the quality of content. In the future, users should be involved in the development process of ACP support tools, and the content should be substantiated by scientific evidence.Trial registrationPROSPERO CRD42020184112; https://tinyurl.com/mruf8b43.

Project description:Policy pointsUsing publicly available Hospital Compare and Medicare data, we found a substantial range of hospital-level performance on quality, expenditure, and value measures for 4 common reasons for admission. Hospitals' ability to consistently deliver high-quality, low-cost care varied across the different reasons for admission. With the exception of coronary artery bypass grafting, hospitals that provided the highest-value care had more beds and a larger average daily census than those providing the lowest-value care. Transparent data like those we present can empower patients to compare hospital performance, make better-informed treatment decisions, and decide where to obtain care for particular health care problems.ContextIn the United States, the transition from volume to value dominates discussions of health care reform. While shared decision making might help patients determine whether to get care, transparency in procedure- and hospital-specific value measures would help them determine where to get care.MethodsUsing Hospital Compare and Medicare expenditure data, we constructed a hospital-level measure of value from a numerator composed of quality-of-care measures (satisfaction, use of timely and effective care, and avoidance of harms) and a denominator composed of risk-adjusted 30-day episode-of-care expenditures for acute myocardial infarction (1,900 hospitals), coronary artery bypass grafting (884 hospitals), colectomy (1,252 hospitals), and hip replacement surgery (1,243 hospitals).FindingsWe found substantial variation in aggregate measures of quality, cost, and value at the hospital level. Value calculation provided additional richness when compared to assessment based on quality or cost alone: about 50% of hospitals in an extreme quality- (and about 65% more in an extreme cost-) quintile were in the same extreme value quintile. With the exception of coronary artery bypass grafting, higher-value hospitals were larger and had a higher average daily census than lower-value hospitals, but were no more likely to be accredited by the Joint Commission or to have a residency program accredited by the American Council of Graduate Medical Education.ConclusionsWhile future efforts to compose value measures will certainly be modified and expanded to examine other reasons for admission, the construct that we present could allow patients to transparently compare procedure- and hospital-specific quality, spending, and value and empower them to decide where to obtain care.

Dataset Information

From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices.

Publications

From What to How: An Initial Review of Publicly Available AI Ethics Tools, Methods and Research to Translate Principles into Practices.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets