Your posts betray you: Detecting influencer-generated sponsored posts by finding the right clues

Shen, Rong-Ping; Liu, Dun; Wu, Xuan; Zhang, Mingyue

doi:10.1016/j.im.2022.103719

Cited by 9 publications

(1 citation statement)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The proper disclosure of ads is a legal requirement. Therefore, this task is fundamental for guaranteeing transparency in social media, protecting consumers from misleading or harmful ads, and ensuring legal compliance (Fatima 2023;Shen et al 2022;Zarei et al 2020;Bertaglia et al 2023;Ershov and Mitchell 2020). Yet, due to limited API access and costly data annotation, data for developing machine learning solutions for detecting such content is scarce.…”

Section: Introductionmentioning

confidence: 99%

InstaSynth: Opportunities and Challenges in Generating Synthetic Instagram Data with ChatGPT for Sponsored Content Detection

Bertaglia,

Heisig,

Kaushal

et al. 2024

ICWSM

View full text Add to dashboard Cite

Large Language Models (LLMs) raise concerns about lowering the cost of generating texts that could be used for unethical or illegal purposes, especially on social media. This paper investigates the promise of such models to help enforce legal requirements related to the disclosure of sponsored content online. We investigate the use of LLMs for generating synthetic Instagram captions with two objectives: The first objective (fidelity) is to produce realistic synthetic datasets. For this, we implement content-level and network-level metrics to assess whether synthetic captions are realistic. The second objective (utility) is to create synthetic data useful for sponsored content detection. For this, we evaluate the effectiveness of the generated synthetic data for training classifiers to identify undisclosed advertisements on Instagram. Our investigations show that the objectives of fidelity and utility may conflict and that prompt engineering is a useful but insufficient strategy. Additionally, we find that while individual synthetic posts may appear realistic, collectively they lack diversity, topic connectivity, and realistic user interaction patterns.

show abstract

Section: Introductionmentioning

confidence: 99%

InstaSynth: Opportunities and Challenges in Generating Synthetic Instagram Data with ChatGPT for Sponsored Content Detection

Bertaglia,

Heisig,

Kaushal

et al. 2024

ICWSM

View full text Add to dashboard Cite

show abstract

How real is real enough? Unveiling the diverse power of generative AI‐enabled virtual influencers and the dynamics of human responses

Sorosrungruang,

Ameen,

Hackley

2024

Psychology and Marketing

View full text Add to dashboard Cite

Virtual influencers in various forms are capturing a growing share of ad spend from human influencers. Followers respond to virtual influencers much as if they were human, with engagement rates and measures of trust and source credibility that rival their human counterparts. However, there is an acute need for more nuanced understanding of the differential characteristics of user engagement with human influencers and the many emerging forms, types and interactional characteristics of virtual influencers. We conduct an exploratory study of three parts. First, through an indepth review of the existing literature, we delineate the implications of unsettled taxonomies of virtual influencers by function and form, and we outline a revised typology. Second, our secondary review of virtual influencer literature, trade, and industry sources conceptualizes divergent factors influencing the persuasive capability of human and virtual influencers while identifying intersecting research themes. From this synthesis we induce suggestions for future research and practice. Finally, we assess, refine and adjust our framework through depth interviews with leading expert practitioners to generate six key findings to guide future researched‐backed virtual influencer practice and research.

show abstract

Apple doesn’t fall far from the tree: Effect of extrinsic factors of online reviews on predicting useless reviews

Choi

2024

Electron Commer Res

View full text Add to dashboard Cite

Your posts betray you: Detecting influencer-generated sponsored posts by finding the right clues

Cited by 9 publications

References 61 publications

InstaSynth: Opportunities and Challenges in Generating Synthetic Instagram Data with ChatGPT for Sponsored Content Detection

InstaSynth: Opportunities and Challenges in Generating Synthetic Instagram Data with ChatGPT for Sponsored Content Detection

How real is real enough? Unveiling the diverse power of generative AI‐enabled virtual influencers and the dynamics of human responses

Apple doesn’t fall far from the tree: Effect of extrinsic factors of online reviews on predicting useless reviews

Contact Info

Product

Resources

About