Equity and Value Alignment in Digital Technologies

The field of digital technologies is moving towards a greater emphasis on equity and value alignment. Researchers are exploring the impact of digital infrastructures on labor markets and social inequalities, highlighting the need for more localized interventions to ensure equitable access to technology. There is also a growing interest in measuring and operationalizing human values in digital systems, including the development of frameworks for measuring value expressions in social media posts and the adaptation of metrics for measuring aspirations in contexts of socio-political instability. Furthermore, scholars are investigating the intersection of AI and justice, with a focus on developing principle-based design frameworks for human services AI and reforming AI model training paradigms to prioritize intrinsic, identity-based development. Notable papers in this area include:

  • A study on measuring value expressions in social media posts, which introduces a framework for measuring Schwartz value expressions and constructs a personalization architecture for classifying value expressions.
  • A paper on operationalizing justice in human services AI, which conducts a mixed-methods analysis of child welfare policy and identifies a range of functional definitions of justice.
  • A call to reform AI model training paradigms, which proposes a paradigm shift from model training to model raising and identifies key components for redesigning the training corpus.

Sources

Local Technological Access, Income Disparities, and Job-Seeking in the United States Since 2010

Measuring Value Expressions in Social Media Posts

Hope, Aspirations, and the Impact of LLMs on Female Programming Learners in Afghanistan

Operationalizing Justice: Towards the Development of a Principle Based Design Framework for Human Services AI

From Model Training to Model Raising - A call to reform AI model training paradigms from post-hoc alignment to intrinsic, identity-based development

From Model Training to Model Raising -- A call to reform AI model training paradigms from post-hoc alignment to intrinsic, identity-based development

Built with on top of