Towards a Programmable Humanizing AI through Scalable Stance-Directed Architecture


ÇETİNKAYA Y. M., Lee Y., KÜLAH E., TOROSLU İ. H., Cowan M. A., Davulcu H.

IEEE Internet Computing, vol.28, no.5, pp.20-27, 2024 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 28 Issue: 5
  • Publication Date: 2024
  • Doi Number: 10.1109/mic.2024.3450090
  • Journal Name: IEEE Internet Computing
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, PASCAL, Aerospace Database, Applied Science & Technology Source, Communication Abstracts, Compendex, Computer & Applied Sciences, INSPEC, Metadex, Civil Engineering Abstracts
  • Page Numbers: pp.20-27
  • Keywords: language generation, language models, sentiment analysis, social networking, Twitter, web text analysis
  • Middle East Technical University Affiliated: Yes

Abstract

The rise of harmful online content underscores the urgent need for AI systems to effectively detect, filter those, and foster safer and healthier communication. This article introduces a novel approach to mitigate toxic content generation propensities of Large Language Models (LLMs) by fine-tuning them with a programmable stance-directed focus on core human values and common good. We propose a streamlined keyword coding and processing pipeline to generate weakly labeled data to train AI models that can avoid toxicity and champion civil discourse. We also developed a toxicity classifier and an Aspect-based Sentiment Analysis (ABSA) model to assess and control the effectiveness of a humanizing AI model. We evaluate the proposed pipeline using a contentious real-world Twitter dataset on U.S. race relations. Our approach successfully curbs the toxic content generation propensity of an unrestricted LLM by a significant 85%.