Machine Learning Technologies and Their Inherent Human Rights Issues in Criminal Justice Contexts

Grace, Jamie

doi:10.2139/ssrn.3487454

Cited by 7 publications

(5 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Depending on the jurisdiction, there may be an orthogonal set of laws that aims to ensure the rights of the users of models created from the data. A number of prior works, particularly from a U.S.-centric perspective, have connected ML to legal frameworks for human rights, especially anti-discrimination [47,48,64,70,72,140,142,146]. These works often focus on the difficulties of constraining algorithmic discrimination in many contexts, proposing alternative legal frameworks that would allow for more regulatory enforcement of algorithmic bias.…”

Section: Legal Context: Rights and Regulationsmentioning

confidence: 99%

Data Governance in the Age of Large-Scale Data-Driven Language Technology

Jernite,

Nguyen,

Biderman

et al. 2022

Preprint

View full text Add to dashboard Cite

The recent emergence and adoption of Machine Learning technology, and specifically of Large Language Models, has drawn attention to the need for systematic and transparent management of language data. This work proposes an approach to global language data governance that attempts to organize data management amongst stakeholders, values, and rights. Our proposal is informed by prior work on distributed governance that accounts for human values and grounded by an international research collaboration that brings together researchers and practitioners from 60 countries. The framework we present is a multi-party international governance structure focused on language data, and incorporating technical and organizational tools needed to support its work.

show abstract

Section: Legal Context: Rights and Regulationsmentioning

confidence: 99%

Data Governance in the Age of Large-Scale Data-Driven Language Technology

Jernite,

Nguyen,

Biderman

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…However, what is sufficient for one user might be insufficient for another as their focus is on a different type of interpretability, or they may have different background knowledge (Doran et al 2017;Hirsch et al 2017). Transparency and interpretability can help to uncover any misbehavior of the algorithms such as inequality, unfairness, and even discrimination and racism (Grace 2019;Hübner 2021).…”

Section: Machine Learning From An End-user Perspectivementioning

confidence: 99%

‘Why are the Sales Forecasts so low?’ Socio-Technical Challenges of Using Machine Learning for Forecasting Sales in a Bakery

Fries

Ludwig

2022

Comput Supported Coop Work

View full text Add to dashboard Cite

Artificial intelligence and the underlying machine learning (ML) methods are increasingly finding their way into our working world. One of these areas is sales planning, where machine learning is used to leverage a variety of different input parameters such as prices, promotions, or the weather, to forecast sales, and therefore directly affects the production of products and goods. To satisfy the goal of environmental sustainability as well as address short shelf life, the food industry represents an interesting application field for the use of ML for optimizing sales planning. Within this paper, we will examine the design, and especially the application, of ML methods in the food industry and show the current challenges that exist in the use of such concepts and technologies from the end-user’s point of view. Our study of a smaller bakery company shows that there are enormous challenges in setting up the appropriate infrastructure and processes for the implementation of ML, that the output quality of ML processes does not always match the perceived result quality, and that trust in the functioning of the algorithms is the most important criterion for using ML processes in practice.

show abstract

“…For example, governments are currently attempting to steer how facial recognition technologies and algorithmic biases are used before they become ubiquitous (e.g. Grace, 2019). However, governments lack the insights into how these technologies will impact their own structures and activities.…”

Section: Box 15 Oecd's Work On Emerging Technologies and Regulationmentioning

confidence: 99%