Defining and Designing an Ethical Approach to Generative Artificial Intelligence in Text-to-Image Modeling

Loading...
Thumbnail Image
Penn collection
The Wharton School::Wharton Undergraduate Research::Joseph Wharton Scholars
Degree type
Discipline
Business
Philosophy
Subject
Artificial Intelligence
Ethics
AI
Business
Equality
Discrimination
Stereotype
Funder
Grant number
Copyright date
2024
Distributor
Related resources
Author
Goodman, Drake
Contributor
Sepinwall, Amy
Abstract

This paper addresses the ethical concerns that generative artificial intelligence (AI) in text-to-image modeling poses, specifically in protecting social equality and against discrimination. It first defines AI and explains the focus on generative AI. It then discusses the emergence of generative AI modeling, as well as prominent players in the generative AI text-to-image space. After explaining different ethical questions that have arisen in response to the rapid deployment of the technology, this paper establishes an ethical claim for why AI developers need to design these models to protect social equality and reduce stereotypes. This occurs in two ways. The first is explaining how the biases and stereotypes present in generative AI differ from the world before this technology existed. The second is establishing why amplifying stereotypes and biases is wrong, as well as why generative AI developers specifically have a moral obligation to protect social equality.

Advisor
Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
2024
Volume number
Issue number
Publisher
Publisher DOI
Journal Issue
Comments
Recommended citation
Collection