Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Persuasion for Good: Towards a Personalized Persuasive Dialogue System for Social Good (1906.06725v2)

Published 16 Jun 2019 in cs.CL, cs.AI, and cs.CY
Persuasion for Good: Towards a Personalized Persuasive Dialogue System for Social Good

Abstract: Developing intelligent persuasive conversational agents to change people's opinions and actions for social good is the frontier in advancing the ethical development of automated dialogue systems. To do so, the first step is to understand the intricate organization of strategic disclosures and appeals employed in human persuasion conversations. We designed an online persuasion task where one participant was asked to persuade the other to donate to a specific charity. We collected a large dataset with 1,017 dialogues and annotated emerging persuasion strategies from a subset. Based on the annotation, we built a baseline classifier with context information and sentence-level features to predict the 10 persuasion strategies used in the corpus. Furthermore, to develop an understanding of personalized persuasion processes, we analyzed the relationships between individuals' demographic and psychological backgrounds including personality, morality, value systems, and their willingness for donation. Then, we analyzed which types of persuasion strategies led to a greater amount of donation depending on the individuals' personal backgrounds. This work lays the ground for developing a personalized persuasive dialogue system.

Overview of "Persuasion for Good: Towards a Personalized Persuasive Dialogue System for Social Good"

The paper "Persuasion for Good: Towards a Personalized Persuasive Dialogue System for Social Good" is a comprehensive paper on developing persuasive dialogue systems aimed at promoting social good, specifically by encouraging individuals to donate to charitable causes. The research addresses the need for intelligent conversational agents that can personalize persuasive strategies based on user demographics and psychological profiles, thereby enhancing the effectiveness of persuasion in dialogues.

Contribution and Methodology

The authors designed an online task where participants engaged in text-based dialogues to persuade counterparts to donate to a charity, "Save the Children". They collected 1,017 dialogues, annotating a subset for various emerging persuasive strategies. The dataset not only encapsulates diverse persuasive tactics but also integrates demographic and psychological data, offering a multi-layered perspective on persuasion.

To classify these strategies, the authors developed a hybrid Recurrent-Convolutional Neural Network (RCNN) model incorporating sentence embeddings, context information, and sentence-level features. Notably, their system distinguished ten persuasion strategies, such as logical and emotional appeals, credibility, and foot-in-the-door techniques. This robust classification process is a foundational step towards automating personalized persuasion.

Findings

Key findings reveal that certain persuasion strategies, such as providing donation information, positively correlate with higher donation rates. Moreover, the effectiveness of persuasive tactics significantly varies with the persuadee's personal background. For instance, strategies like emotion appeal were more effective with extroverted individuals, demonstrating the importance of aligning persuasive strategies with psychological traits.

The paper also highlights the role of user demographics, such as age and decision-making style, in influencing donation outcomes. Older participants or those with a rational decision-making style showed a greater propensity to donate. This insight is crucial for tailoring persuasion strategies based on user profiles.

Implications and Future Directions

The research paves the way for more sophisticated, personalized persuasive systems that could revolutionize how AI interacts with humans in contexts requiring behavior change, such as health and environmental conservation. The integration of psychological and demographic data into the persuasion model exemplifies a forward-thinking approach that could significantly improve the efficacy of conversational agents.

Moving forward, the paper suggests enhancing the dialogue system by refining its ability to extract and process contextual information of conversations. Future research could extend this work by applying it to other domains where persuasion plays a critical role, ensuring the dialogue systems are ethically designed to support user autonomy and decision-making.

In conclusion, the paper provides a compelling case for the development of personalized persuasive dialogues, illustrating how nuanced, data-driven approaches can enhance the effectiveness of AI in promoting social good. The insights on the interaction between strategy efficacy and user traits are invaluable for practitioners aiming to leverage AI for behavior change interventions.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (7)
  1. Xuewei Wang (14 papers)
  2. Weiyan Shi (41 papers)
  3. Richard Kim (7 papers)
  4. Yoojung Oh (2 papers)
  5. Sijia Yang (18 papers)
  6. Jingwen Zhang (54 papers)
  7. Zhou Yu (206 papers)
Citations (240)