What to explain when explaining is difficult. An interdisciplinary primer on XAI and meaningful information in automated decision-making

Asghari, Hadi; Birner, Nadine; Burchardt, Aljoscha; Dicks, Daniela; Judith, Faßbender,; Feldhus, Nils; Hewett, Freya; Hofmann, Vincent; Kettemann, Matthias C.; Schulz, Wolfgang A.; Simon, Judith; Stolberg-Larsen, Jakob; Züger, Theresa

doi:10.5281/zenodo.6375784

Cited by 2 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…XAI is thus not merely a technical challenge. There are always social components that affect how explanations are designed, e.g., for whom the explanations are made, what specific needs the users have, all the way to the question of what educational offerings are necessary to be able to understand the explanations given ( 37 , 38 ).…”

Section: Discussionmentioning

confidence: 99%

“Nothing works without the doctor:” Physicians’ perception of clinical decision-making and artificial intelligence

et al. 2022

Self Cite

View full text Add to dashboard Cite

IntroductionArtificial intelligence–driven decision support systems (AI–DSS) have the potential to help physicians analyze data and facilitate the search for a correct diagnosis or suitable intervention. The potential of such systems is often emphasized. However, implementation in clinical practice deserves continuous attention. This article aims to shed light on the needs and challenges arising from the use of AI-DSS from physicians’ perspectives.MethodsThe basis for this study is a qualitative content analysis of expert interviews with experienced nephrologists after testing an AI-DSS in a straightforward usage scenario.ResultsThe results provide insights on the basics of clinical decision-making, expected challenges when using AI-DSS as well as a reflection on the test run.DiscussionWhile we can confirm the somewhat expectable demand for better explainability and control, other insights highlight the need to uphold classical strengths of the medical profession when using AI-DSS as well as the importance of broadening the view of AI-related challenges to the clinical environment, especially during treatment. Our results stress the necessity for adjusting AI-DSS to shared decision-making. We conclude that explainability must be context-specific while fostering meaningful interaction with the systems available.

show abstract

Section: Discussionmentioning

confidence: 99%

“Nothing works without the doctor:” Physicians’ perception of clinical decision-making and artificial intelligence

et al. 2022

Self Cite

View full text Add to dashboard Cite

show abstract

“…This does not necessarily have to include complete disclosure of the entire technical functioning of the system. (Asghari et al, 2021) AI systems treat explanations as static products, calculated once and for all. In reality, an adequate explanation is more of a process than a product, requiring dynamic, iterative refinement among multiple agents.…”

Section: Commentsmentioning

confidence: 99%

Trustworthy AI Alone Is Not Enough.

Pérez y Madrid,

Wright

2023

View full text Add to dashboard Cite

Recent advances in AI through machine learning (ML) 1 and deep learning (DL) 2 techniques have arrived with a lot of momentum in society. In the 2010s there were several developments. The large amount of available data and the development of advanced model fitting algorithms based on artificial neural networks (ANN) 3 enabled the creation of advanced pattern recognition systems. First, the coveted task of handwritten digit recognition was achieved. The next stage was the recognition of shapes in images and faces, and from then its usefulness in other fields began to be tested, from industrial applications to others related to human behaviour.The great hope was and still is the possibility of solving problems considered reserved for human intelligence. During these years it has been possible to create models which we classify in two categories: GLOSSARY Artificial neural networks (ANNs): usually simply called neural networks (NN) or neural nets, are computing systems inspired by the biological neural networks that constitute animal brains (see also https://en.wikipedia.org/ wiki/Artificial_neural_network). Black box: A black box is a system that can only be known in terms of its inputs and outputs without knowing its inner workings. Its mode of operation is opaque, hence the name black. AI algorithms using artificial neural networks typically require a huge number of parameters. This makes human understanding of their operation impossible. For this reason, such algorithms are considered black boxes. Data Science: Data science is an interdisciplinary field focused on extracting knowledge from large data sets and applying that knowledge and insights to solve problems in a wide range of application domains. The field includes preparing data for analysis, framing data science problems, analysing data, developing data-driven solutions and presenting results to inform high-level decisions in a wide range of application domains. As such, it incorporates knowledge of computer science, statistics, mathematics, data visualisation, information visualisation, data integration, graphic design, complex systems, communication and business (See also https://en.wikipedia.org/wiki/Data_ science). Deep Learning (DL): Deep learning is part of a broader family of machine learning methods, which is based on artificial neural networks. The adjective "deep" in deep learning refers to the use of multiple layers in the network. Generative AI: is a type of artificial intelligence system capable of generating text, images or other media in response to instructions. Generative AI models learn the patterns and structure of their input training data and then generate new data with similar characteristics. Machine learning (ML): is a branch of AI that allows machines to learn from large amounts of data. Metadata: data that describe other data, e.g. their information content. They are analogous to indexes used in a library, such as author, title, publisher, date, etc. In the case of Spotify this could be singer, composer, style, etc.

show abstract

What to explain when explaining is difficult. An interdisciplinary primer on XAI and meaningful information in automated decision-making

Cited by 2 publications

References 0 publications

“Nothing works without the doctor:” Physicians’ perception of clinical decision-making and artificial intelligence

“Nothing works without the doctor:” Physicians’ perception of clinical decision-making and artificial intelligence

Trustworthy AI Alone Is Not Enough.

Contact Info

Product

Resources

About