Can AI Moderate Online Communities?

Research output: Working paper › Preprint › Research

Standard

Can AI Moderate Online Communities? / Axelsen, Henrik Bjørn; Jensen, Johannes Rude; Axelsen, Sebastian ; Licht, Valdemar ; Ross, Omry.

arXiv preprint, 2023.

Research output: Working paper › Preprint › Research

Harvard

Axelsen, HB, Jensen, JR, Axelsen, S, Licht, V & Ross, O 2023 'Can AI Moderate Online Communities?' arXiv preprint. <https://arxiv.org/abs/2306.05122>

APA

Axelsen, H. B., Jensen, J. R., Axelsen, S., Licht, V., & Ross, O. (2023). Can AI Moderate Online Communities? arXiv preprint. https://arxiv.org/abs/2306.05122

Vancouver

Axelsen HB, Jensen JR, Axelsen S, Licht V, Ross O. Can AI Moderate Online Communities? arXiv preprint. 2023.

Author

Axelsen, Henrik Bjørn ; Jensen, Johannes Rude ; Axelsen, Sebastian ; Licht, Valdemar ; Ross, Omry. / Can AI Moderate Online Communities?. arXiv preprint, 2023.

Bibtex

@techreport{c0c7381f12e84a6183721ab5d6c366b9,
title = "Can AI Moderate Online Communities?",
abstract = "The task of cultivating healthy communication in online communities becomes increasingly urgent, as gaming and social media experiences become progressively more immersive and life-like. We approach the challenge of moderating online communities by training student models using a large language model (LLM). We use zero-shot learning models to distill and expand datasets followed by a few-shot learning and a fine-tuning approach, leveraging open-access generative pre-trained transformer models (GPT) from OpenAI. Our preliminary findings suggest, that when properly trained, LLMs can excel in identifying actor intentions, moderating toxic comments, and rewarding positive contributions. The student models perform above-expectation in non-contextual assignments such as identifying classically toxic behavior and perform sufficiently on contextual assignments such as identifying positive contributions to online discourse. Further, using open-access models like OpenAI's GPT we experience a step-change in the development process for what has historically been a complex modeling task. We contribute to the information system (IS) discourse with a rapid development framework on the application of generative AI in content online moderation and management of culture in decentralized, pseudonymous communities by providing a sample model suite of industrial-ready generative AI models based on open-access LLMs.",
author = "Axelsen, {Henrik Bj{\o}rn} and Jensen, {Johannes Rude} and Sebastian Axelsen and Valdemar Licht and Omry Ross",
year = "2023",
language = "English",
publisher = "arXiv preprint",
type = "WorkingPaper",
institution = "arXiv preprint",

}

RIS

TY - UNPB

T1 - Can AI Moderate Online Communities?

AU - Axelsen, Henrik Bjørn

AU - Jensen, Johannes Rude

AU - Axelsen, Sebastian

AU - Licht, Valdemar

AU - Ross, Omry

PY - 2023

Y1 - 2023

N2 - The task of cultivating healthy communication in online communities becomes increasingly urgent, as gaming and social media experiences become progressively more immersive and life-like. We approach the challenge of moderating online communities by training student models using a large language model (LLM). We use zero-shot learning models to distill and expand datasets followed by a few-shot learning and a fine-tuning approach, leveraging open-access generative pre-trained transformer models (GPT) from OpenAI. Our preliminary findings suggest, that when properly trained, LLMs can excel in identifying actor intentions, moderating toxic comments, and rewarding positive contributions. The student models perform above-expectation in non-contextual assignments such as identifying classically toxic behavior and perform sufficiently on contextual assignments such as identifying positive contributions to online discourse. Further, using open-access models like OpenAI's GPT we experience a step-change in the development process for what has historically been a complex modeling task. We contribute to the information system (IS) discourse with a rapid development framework on the application of generative AI in content online moderation and management of culture in decentralized, pseudonymous communities by providing a sample model suite of industrial-ready generative AI models based on open-access LLMs.

AB - The task of cultivating healthy communication in online communities becomes increasingly urgent, as gaming and social media experiences become progressively more immersive and life-like. We approach the challenge of moderating online communities by training student models using a large language model (LLM). We use zero-shot learning models to distill and expand datasets followed by a few-shot learning and a fine-tuning approach, leveraging open-access generative pre-trained transformer models (GPT) from OpenAI. Our preliminary findings suggest, that when properly trained, LLMs can excel in identifying actor intentions, moderating toxic comments, and rewarding positive contributions. The student models perform above-expectation in non-contextual assignments such as identifying classically toxic behavior and perform sufficiently on contextual assignments such as identifying positive contributions to online discourse. Further, using open-access models like OpenAI's GPT we experience a step-change in the development process for what has historically been a complex modeling task. We contribute to the information system (IS) discourse with a rapid development framework on the application of generative AI in content online moderation and management of culture in decentralized, pseudonymous communities by providing a sample model suite of industrial-ready generative AI models based on open-access LLMs.

M3 - Preprint

BT - Can AI Moderate Online Communities?

PB - arXiv preprint

ER -

ID: 384260551