Making Intelligence: Ethical Values in IQ and ML Benchmarks
arXiv:2209.00692 · doi:10.1145/3593013.3593996
Abstract
In recent years, ML researchers have wrestled with defining and improving machine learning (ML) benchmarks and datasets. In parallel, some have trained a critical lens on the ethics of dataset creation and ML research. In this position paper, we highlight the entanglement of ethics with seemingly ``technical'' or ``scientific'' decisions about the design of ML benchmarks. Our starting point is the existence of multiple overlooked structural similarities between human intelligence benchmarks and ML benchmarks. Both types of benchmarks set standards for describing, evaluating, and comparing performance on tasks relevant to intelligence -- standards that many scholars of human intelligence have long recognized as value-laden. We use perspectives from feminist philosophy of science on IQ benchmarks and thick concepts in social science to argue that values need to be considered and documented when creating ML benchmarks. It is neither possible nor desirable to avoid this choice by creating value-neutral benchmarks. Finally, we outline practical recommendations for ML benchmark research ethics and ethics review.
FAccT 2023, June 12 to 15, 2023, Chicago, IL, USA
References in corpus (10)
- Do Datasets Have Politics? Disciplinary Values in Computer Vision Dataset Development
- A Large-scale Study of Representation Learning with the Visual Task Adaptation Benchmark
- Algorithmic Monoculture and Social Welfare
- Holistic Evaluation of Language Models
- CrowdWorkSheets: Accounting for Individual and Collective Identities Underlying Crowdsourced Dataset Annotation
- Institutionalising Ethics in AI through Broader Impact Requirements
- Overcoming Failures of Imagination in AI Infused System Development and Deployment
- Should attention be all we need? The epistemic and ethical implications of unification in machine learning
- A Validity Perspective on Evaluating the Justified Use of Data-driven Decision-making Algorithms
- Evaluation for Change