Browsing by Author "Paakki, Henna"
Now showing 1 - 7 of 7
- Results Per Page
- Sort Options
- Asymmetric Conversational Strategies - Methods for Detecting Manipulative Online Trolling
School of Science | Doctoral dissertation (article-based)(2024) Paakki, HennaThis dissertation examines harmful forms of deceptive manipulative online trolling from a socio-technical perspective. Adopting a transdisciplinary approach grounded in computational linguistics, the work utilizes qualitative digital Conversation Analysis (CA), statistical analysis, Machine Learning (ML) and Natural Language Processing (NLP) methodologies. The focus is on effective identification of conversational trolling: qualitatively defining its central characteristics and developing computational trolling detection based on operationalizable conversational features. Trolling strategies are increasingly used for online harassment and manipulation. This highlights the importance of developing efficient computational trolling detection to protect deliberative online discussion and democratic decision-making. The detection of deceptive behaviors like trolling is a challenging task that earlier research has not managed to accurately accomplish. This is due to the elusive and changing strategies employed by trolls to stir disruption. Approaches to computational trolling detection have traditionally investigated networks of bots, or the valence and content of alone-standing messages separated from their conversational context. However, these methods manage to capture only a fraction of harmful trolling behaviors. Computer-Mediated Communication research has shed light on the context-dependent and interactive nature of trolling, and common trolling styles found online. Based on these insights, this work investigates conversational trolling on several types of asynchronous online forums, collecting a novel dataset. This enables a qualitative depiction of turn-by-turn conversational strategies used in trolling, and their computational analysis. This work shows that while interest-based conversations commonly attract overt trolling styles, political and societal discussions are usually targeted with covert styles. However, they all utilize similar manipulative conversational strategies. These include the use of asymmetric responses (ignoring, challenging, and mismatching) that violate common conversational norms. This dissertation introduces a novel approach to computationally analyzing the dynamics of asynchronous conversations by drawing from digital CA. Together with NLP methods, this allows the extraction of conversational features like asymmetries to train ML models for trolling detection. The work presents a trolling detection model that surpasses earlier models in performance, and suggests a process for detecting deceptive manipulative online behaviors. Due to the challenges related to judging the intentionality behind online behaviors, I suggest an intent-agnostic detection approach based on observable behaviors in interaction. These include repeated violations of common conversational norms, which characterize deceptive manipulative behaviors like trolling. - Conceptualizing Human-Computer Intersubjectivity to Develop Computational Humor
A4 Artikkeli konferenssijulkaisussa(2021-09-14) Paakki, HennaHumor – as well as language in general – is by nature social and tied to a context. To better engage with context, computational humor could draw inspiration from the concept of intersubjectivity: the sharing of perspectives. This paper focuses on discussing the possible advantages of utilizing the concept of intersubjectivity to contextualize computational humor. Intersubjectivity in humor generation system design is discussed as a possible means of evaluation of the creative product, as well as a potential approach to generating more impressive humoristic content. Firstly, evaluation of computational humor has been wanting for more effective and versatile methods. To this problem, an implementation of sharing perspectives between the system and its users offers a viable solution. Secondly, approaches to humor generation are contrasted with interactive dialogue systems, to analyze how they contextualize humor. The comparisons show that well defined interactive design and evaluation methods that enable perspective sharing between the producer and the press would greatly benefit humor-generating systems. The final section theorizes on the possible foundations for modeling intersubjectivity in computational humor. - Detecting covert disruptive behavior in online interaction by analyzing conversational features and norm violations
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2024-01-29) Paakki, Henna; Vepsäläinen, Heidi; Salovaara, Antti; Zafar, BushraDisruptive behavior is a prevalent threat to constructive online engagement. Covert behaviors, such as trolling, are especially challenging to detect automatically, because they utilize deceptive strategies to manipulate conversation. We illustrate a novel approach to their detection: analyzing conversational structures instead of focusing only on messages in isolation. Building on conversation analysis, we demonstrate that (1) conversational actions and their norms provide concepts for a deeper understanding of covert disruption, and that (2) machine learning, natural language processing and structural analysis of conversation can complement message-level features to create models that surpass earlier approaches to trolling detection. Our models, developed for detecting overt (aggression) as well as covert (trolling) behaviors using prior studies' message-level features and new conversational action features, achieved high accuracies (0.90 and 0.92, respectively). The findings offer a theoretically grounded approach to computationally analyzing social media interaction and novel methods for effectively detecting covert disruptive conversations online. - Disruptive online communication: How asymmetric trolling-like response strategies steer conversation off the track
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2021-06) Paakki, Henna; Vepsäläinen, Heidi; Salovaara, AnttiInternet trolling, a form of antisocial online behavior, is a serious problem plaguing social media. Skillful trolls can lure entire communities into degenerative and polarized discussions that continue endlessly. From analysis of data gathered in accordance with established classifications of trolling-like behavior, the paper presents a conversation analysis of trolling-like interaction strategies that disrupt online discussions. The authors argue that troll-like users exploit other users’ desire for common grounding – i.e., joint maintenance of mutual understanding and seeking of conversational closure – by responding asymmetrically. Their responses to others deviate from expectations for typical paired actions in turn-taking. These asymmetries, described through examples of three such behaviors – ignoring, mismatching, and challenging – lead to dissatisfactory interactions, in that they subvert other users’ desire for clarification and explanation of contra-normative social behavior. By avoiding clarifi- cations, troll-like users easily capture unsuspecting users’ attention and manage to prolong futile conversations interminably. Through the analysis, the paper connects trolling-like asymmetric response strategies with concrete data and addresses the implications of this nonconformist behavior for common grounding in social-media venues. - Do Online Trolling Strategies Differ in Political and Interest Forums: Early Results
A4 Artikkeli konferenssijulkaisussa(2020-10-19) Paakki, Henna; Salovaara, Antti; Vepsäläinen, HeidiThis study compares the effectiveness of different trolling strategies in two online contexts: politically oriented forums that address issues like global warming, and interest-based forums that deal with people’s personal interests. Based on previous research, we consider trolling as context-bound and suggest that relevance theory and common grounding theory can explain why people may attend and react to certain types of troll posts in one forum, but pay scant attention to them in another. We postulate two hypotheses on how successful (i.e., disruptive) trolling varies according to context: that trolls’ messaging strategies appear in different frequencies in political and interest forums (H1), and that context-matching strategies also produce longer futile conversations (H2). Using Hardaker’s categorization of trolling strategies on a covert–overt continuum, our statistical analysis on a dataset of 49 online conversations verified H1: in political forums covert strategies were more common than overt ones; ininterest forums the opposite was the case. Regarding H2 our results were inconclusive. However, the results motivate further research on this phenomenon with larger datasets. - How Finnish News Actors Perceive and Combat Visual Disinformation
Perustieteiden korkeakoulu | Master's thesis(2024-06-18) Lares, SanniThe rapid development of artificial intelligence has allowed large masses of people to create ever more believable images and videos without the need for technical expertise. In Finland, trust in the media is high. At its worst, when disinformation gets published in the news media, it has the potential to reach many people who fail to be critical towards it. This thesis examines the understanding and readiness of the Finnish news media field towards visual disinformation. Although especially AI-based visual disinformation occurs in the public debate, there is hardly any scientific research on visual disinformation in the context of news media. This work aims to increase the understanding of visual disinformation as a phenomenon and provide Finnish news media with better tools and practices to tackle it. This study engages qualitative research using the grounded theory method, with 16 semi-structured interviews conducted with six Finnish news media organisations. Based on this work, it can be concluded that the Finnish news media is aware of the existence and threat of visual disinformation, but journalists don’t have sufficient resources, training and capacity to effectively deal with it due to the intensive news cycle, reliance on their professional skills, and lack of knowledge regarding suitable tools available. News media companies have undertaken concrete steps to tackle visual disinformation: media actors offer training, and few Finnish news media organisations have hired fact-checkers to assist in fact-checking, particularly concentrating on contentious war-related news content. This work contributes to a better understanding of the Finnish news media's attitude and readiness to confront visual disinformation. By examining the differences in practices and tools used among specialists and journalists, this work delves into the challenges and opportunities for effectively dealing with visual disinformation in the Finnish news media. - Normativity in English Oral Production in Finland and Japan
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä(2020-10-19) Paakki, HennaThis research examines the effect of normativity on difficulties experienced with English oral production in Finland and Japan. Moyer’s classification of factors influencing second language acquisition (2004) as well as language ideology theory (Garrett, 2010; Milroy, 2007) are used as a framework for an analysis of 56 semi-structured interviews with Finnish and Japanese adult learners of intermediate level English. Self-reported experiences related to speaking English were annotated with appropriate codes and analyzed using content analysis. The results show that normativity related to the English language explains many of the difficulties learners experience with speaking English, and that this normativity is essentially connected to social factors as well as instruction and input factors in language learning.