Reflection machines: increasing meaningful human control over Decision Support Systems

Cornelissen, N. A. J.; Eerdt, R. J. M. van; Schraffenberger, Hanna; Haselager, Pim

doi:10.1007/s10676-022-09645-y

Cited by 6 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In medicine, Topol (2019) has expressed the view that it is unlikely that developments will surpass the level of conditional automation without taking humans out of the loop. There is a risk that humans will end up on or even out of the loop in these automated systems (Cornelissen et al, 2022). Ongoing dialogue between educational professionals, researchers, and developers is needed to keep humans in the loop and evaluate the augmentation perspective's functioning.…”

Section: S IX Le Vel S Of Automati On Of Ai In Educ Ati Onmentioning

confidence: 99%

Towards hybrid human‐AI learning technologies

Molenaar

2022

Euro J of Education

View full text Add to dashboard Cite

Education is a unique area for application of artificial intelligence (AI). In this article, the augmentation perspective and the concept of hybrid intelligence are introduced to frame our thinking about AI in education. The involvement of quadruple helix stakeholders (i.e., researchers, education professionals, entrepreneurs, and policymakers) is necessary to understand the opportunities and challenges of different educational use cases from an integrated point of view. To facilitate a meaningful dialogue, a common language about AI in education is needed. This article outlines elements of such a common language. The detectdiagnose-act framework is used to describe the core functions of AI in education. The six levels of automation model is introduced to develop our thinking about the roles of AI, learners, and teachers in educational arrangements. In this model, the transition of control between teacher and technology is articulated at different levels and related to the augmentation perspective. Finally, the future of AI in education is discussed using self-regulated learning as an example. The proposed common language will help to support a coordinated development of an interdisciplinary dialogue between quadruple helix stakeholders to strengthen meaningful application of AI for learning and teaching.

show abstract

Section: S IX Le Vel S Of Automati On Of Ai In Educ Ati Onmentioning

confidence: 99%

Towards hybrid human‐AI learning technologies

Molenaar

2022

Euro J of Education

View full text Add to dashboard Cite

show abstract

“…The reviewer may then abandon their own edits because the auto-suggestion seems to provide an answer. One way to minimize such risk is the deployment of Reflection Machines (Cornelissen et al, 2022), a decision support system that will compel users to give reasons for accepting or rejecting suggestions from the model.…”

Section: Ethical Considerationsmentioning

confidence: 99%

DuluthNLP at SemEval-2022 Task 7: Classifying Plausible Alternatives with Pre–trained ELECTRA

Akrah¹,

Pedersen²

2022

Proceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)

View full text Add to dashboard Cite

This paper describes the DuluthNLP system that participated in Task 7 of SemEval-2022 on Identifying Plausible Clarifications of Implicit and Underspecified Phrases in Instructional Texts. Given an instructional text with an omitted token, the task requires models to classify or rank the plausibility of potential fillers. To solve the task, we fine-tuned the models BERT, RoBERTa, and ELECTRA on training data where potential fillers are rated for plausibility. This is a challenging problem, as shown by BERT-based models achieving accuracy less than 45%. However, our ELECTRA model with tuned class weights on CrossEntropyLoss achieves an accuracy of 53.3% on the official evaluation test data, which ranks 6 out of the 8 total submissions for Subtask A.

show abstract

“…Future studies could investigate the judgments of investment funds that over-rely on decision support systems, i.e. do not conduct necessary due diligence when using information suggested by algorithms (Cornelissen et al, 2022).…”

Section: Limitationsmentioning

confidence: 99%

Will the use of machines to identify negative activist investments backfire? Experimental evidence

Niszczota¹,

Kaszás²

2022

Preprint

View full text Add to dashboard Cite

Current research does not resolve how people will judge harmful activist investments if machines (machine learning algorithms) are involved in the investment process as advisors, and not the ones “pulling the trigger”. On the one hand, machines might diffuse responsibility for making a socially responsible, but harmful investment. On the other hand, machines could exacerbate the blame that is assigned to the investment fund, which can be penalized for outsourcing part of the decision process to an algorithm. We attempted to resolve this issue experimentally. In our experiment (N = 956), participants judge either a human research team or a machine learning algorithm as the source of advice to the investment team to short-sell a company that they suspect of committing financial fraud. Results suggest that investment funds will be similarly blameworthy for an error regardless of using human or machine intelligence to support their decision to short-sell a company. This finding highlights a novel and relevant circumstance in which reliance on algorithms does not backfire by making the final decision-maker (e.g., an investment fund) more blameworthy. Nor does it lessen the perceived blameworthiness of the final decision-maker, by making algorithms into “electronic scapegoats” for providing well-intended but harmful advice.

show abstract

Reflection machines: increasing meaningful human control over Decision Support Systems

Cited by 6 publications

References 16 publications

Towards hybrid human‐AI learning technologies

Towards hybrid human‐AI learning technologies

DuluthNLP at SemEval-2022 Task 7: Classifying Plausible Alternatives with Pre–trained ELECTRA

Will the use of machines to identify negative activist investments backfire? Experimental evidence

Contact Info

Product

Resources

About