News Feeds Research Articles

Deep Dive into Fake News Detection: Feature-Centric Classification with Ensemble and Deep Learning Methods

The online spread of fake news on various platforms has emerged as a significant concern, posing threats to public opinion, political stability, and the dissemination of reliable information. Researchers have turned to advanced technologies, including machine learning (ML) and deep learning (DL) techniques, to detect and classify fake news to address this issue. This research study explores fake news classification using diverse ML and DL approaches. We utilized a well-known “Fake News” dataset sourced from Kaggle, encompassing a labelled news collection. We implemented diverse ML models, including multinomial naïve bayes (MNB), gaussian naïve bayes (GNB), Bernoulli naïve Bayes (BNB), logistic regression (LR), and passive aggressive classifier (PAC). Additionally, we explored DL models, such as long short-term memory (LSTM), convolutional neural networks (CNN), and CNN-LSTM. We compared the performance of these models based on key evaluation metrics, such as accuracy, precision, recall, and the F1 score. Additionally, we conducted cross-validation and hyperparameter tuning to ensure optimal performance. The results provide valuable insights into the strengths and weaknesses of each model in classifying fake news. We observed that DL models, particularly LSTM and CNN-LSTM, showed better performance compared to traditional ML models. These models achieved higher accuracy and demonstrated robustness in classification tasks. These findings emphasize the potential of DL models to tackle the spread of fake news effectively and highlight the importance of utilizing advanced techniques to address this challenging problem.

Mainstream Media Strategy Embracing Muslim Centennials

The mainstream mass media has recorded its history in the journey of Indonesian journalism. The development and condition of the Covid-19 pandemic that utilizes digitalization technology have forced the mainstream mass media to transform to survive. The mainstream media is still read by Gen Z or centennials even though in the onslaught of social media and the shift of the age of information seekers or readers. Despite using social media and news aggregators as their first information sources, centennials still use mainstream media to clarify dubious facts they found on digital media. This study explores how the mainstream mass media survive and embrace the 100th generation as them. The researcher used an approach with case study methods and data collection techniques using observation, in-depth interviews, and document studies. The research subjects are editors of online mainstream media, namely suaramerdeka.com, krjogja.com, pikiran-rakyat.com, and Republika.co.id. This study discusses the strategic differences between the online mainstream media in the regions and the national media. Three interest findings are the highlight of this study. First, the routine media and approaches used to embrace the centennial generation as a potential audience. Second, the existence of content creators in the ranks of journalists is a new phenomenon that presents pros and cons. Third, the Republika.com approach as a media that serves Muslim audiences to Muslim centennials. Lastly, the use of journalistic language style that is familiar to centennials.
 Keywords: mainstream media, Muslim centennials, content creator, media routine

Stocks and fishery of cod (<i>Gadus macrocephalus, Gadidae</i>) in the northwestern Bering Sea in 1965–2022

For pacific cod in the northwestern Bering Sea, biomass of the summer-fall feeding aggregations in the specified area of their highest density (polygon) is considered as a stock unit. The fish in these aggregations belong to the populations spawn in both local and adjacent areas. The stock values, biological parameters, and annual catches are determined for the period from 1965 to 2022 on the materials collected in research surveys and commercial expeditions, as well as on the data of fishery statistics and cited data. The average length of cod caught by different commercial fishing gears is determined as 42.01 cm for Danish seine, 53.01 cm for bottom trawls, 58.15 cm for midwater trawls, 61.94 cm for longlines. The midwater and bottom trawls used in research surveys are equipped with the rigging for better account of juvenile fish, therefore the average length of cod in their catches is 46.42 and 36.52 cm, respectively. The cod aged from 3 to 12 years are presented in the longline catches, with the modal group of 5 years old, but the age range is wider for bottom trawls, from yearlings to 12–13 years old fish, with a modal group varying from 2 to 5 years old in different years. The parameters of von Bertalanffy equation of linear growth for pacific cod are determined as: L∞ = 111.993 cm, K = 0.159, and t0 = –0.024 and for the equation of weight growth rate: W∞ = 14.006 kg, K = 0.245, and t0 = 1.221. The age of mass maturation is 4.4 years. The instantaneous coefficient of natural mortality, calculated by Tyurin, is age-dependent, with the minimum for 5 years old fish (0.345 year–1), 1.245 year–1 for yearlings, 1.231 year–1 for 11 years old fish, and 1.860 year–1 for 12 years old fish. History of the cod fishery in the northwestern Bering Sea is traced since its start in 1965. The cod fishery was realized with various types of fishing vessels (from small seiners to supertrawlers) and fishing gears, but the bulk of annual catch was landed with longlines (65.9 % in the 2010–2022, whereas only 21.4 % was landed with bottom trawls and 12.7 % with Danish seine). Overfishing happened in 1971, when the annual catch increased to 91,600 t on the background of poor replenishment, after that the catch decreased to 760–1420 t in 1976–1979. Then the stock had recovered and the catch increased from 3890 t in 1980 to 58,480 t in 1986. Next decline to 13,810 t in 2009 was followed by a new growth to the record value of 106.2 . 103 t in 2020, but recently the annual catch decreased again (61.5 . 103 t in 2022). This dynamics corresponded to the stock fluctuations: according to the trawl surveys data, the spawning stock within the polygon ranged between 66.5–282.6 . 103 tons in the 1965–1974, decreased to 24.5 . 103 t in 1977 because of overfishing, gradually recovered to 108.1 . 103 t in 1980, 173.0 . 103 t in 1982, and 561.6 . 103 t in 2000, then declined again to 151.4 . 103 t in 2001 and grew in the last two decades to the maximum values of 1,297.9 . 103 t in 2018. Recently a rapid decline of the stock is observed to 475.0 . 103 t in 2022. The changes of the assessed spawning stock generally precede for two years the CPUE changes due to relatively younger age composition in the catches of research bottom trawl used for the stock calculation, in comparison with the age composition in commercial catches. The stock dynamics within the polygon in the northwestern Bering Sea is driven by variations of the cod feeding migration to this area that correlates with the extension of the coldwater pool at the sea bottom around St. Lawrence Island. From the other hand, the migration depends on the cod abundance on the spawning grounds that could be negatively impacted by the bottom trawl fishery in the spawning season. Since 2002, the fishing pressure was moderate, and fishing mortality did not exceed the fishing exploitation target. However, overfishing is possible in the next 3–4 years, if the fishing mortality goes beyond the target benchmark.

Efektivitas Sarana Temu Kembali Informasi Pada Sistem Media Asset Management Dengan Tinjaun Recall dan Precision di PT. Surya Citra Media Tbk

Purpose Research. In order to provide information about visual news shows needed by the News division staff at PT Surya Citra Media Tbk quickly, precisely, and relevant as needed, the Media Asset Management system which is a means of retrieving information in the News Library at PT Surya Citra Media Tbk, needs to be seen and measured in order to know its effectiveness and ideality in bringing together information that suits the needs. This study aims to determine the effectiveness of the Media Asset Management system in the News Library of PT Surya Citra Media Tbk. Method. The method used is a formative evaluation research method with a quantitative approach. The population in this study is the entire collection of news shows stored in the Backdrop category in 2022. The sampling technique used was purposive sampling. The research sample is 20 news keywords entered into the Media Asset Management system. This research uses Lancaster's Recall and Precision effectiveness theory. Analysis Data. The research data were analyzed using descriptive statistical analysis techniques Results. The results showed an average Recall calculation ratio of 0.45 or 45% and an average Precision calculation ratio of 0.89 or 89%. Based on the nominal Precision level, the Media Asset Management system falls into the third category with a high or effective Precision level. Based on the identification of the comparison between the average Recall and Precision ratios, the Media Asset Management system has not been categorized in an ideal condition with a ratio between the Recall and Precision ratios that are not equally large, namely (4: 9). Conclusions. The Media Asset Management system in the News Library of PT Surya Citra Media Tbk is classified as effective but not yet categorized as being in an ideal stage or condition. Keywords: Effectiveness; Media Asset Management; Precision; Recall

The essential habitat role of a unique coastal inlet for a widely distributed apex predator.

Essential habitats support specific functions for species, such as reproduction, feeding or refuge. For highly mobile aquatic species, identifying essential habitats within the wider distribution range is central to understanding species ecology, and underpinning effective management plans. This study examined the movement and space use patterns of sevengill sharks (Notorynchus cepedianus) in Caleta Valdés (CV), a unique coastal habitat in northern Patagonia, Argentina. Seasonal residency patterns of sharks were evident, with higher detectability in late spring and early summer and lower during autumn and winter. The overlap between the residency patterns of sharks and their prey, elephant seals, suggests that CV functions as a seasonal feeding aggregation site for N. cepedianus. The study also found sexual differences in movement behaviour, with males performing abrupt departures from CV and showing increased roaming with the presence of more sharks, and maximum detection probability at high tide. These movements could be related to different feeding strategies between sexes or mate-searching behaviour, suggesting that CV may also be essential for reproduction. Overall, this study highlights the importance of coastal sites as essential habitats for N. cepedianus and deepens our understanding of the ecological role of this apex predator in marine ecosystems.

Media Coverage of Medical Tourism In Russia (2022)

The paper represents the results of the analysis of publications devoted to the coverage of medical tourism in Russia. The study period is limited from the beginning of the special military operation until the end of 2022. The selection of publications was made in the Yandex news aggregator and the Medialogy database. The article compares the measurements received in the beginning and at the end of the period (February – December 2022). The obtained results are a part of a panel study dedicated to media coverage of the healthcare system in Russia and can be used to adjust the information field by relevant structures.

Using video for fact-checking on Facebook. Analysis of the trend and reach of Ibero-American production (2016-2021)

Social networks –and Facebook in particular– have become an important element of the information diet for millions of people around the world. By using them, the traditional media lose control of the distribution channel for their content, whose reach now depends, firstly, on the relevance criteria established by the networks themselves and, secondly, on the interactions generated by the audience with each publication. Very often research on journalism has tackled the issue of reach and how efficient fact-checkers are. To find answer to the sociodemographic features of their audiences or the characteristics of their posts are explored. However, factors such as the influence of the algorithms which choose the content users are shown on the social networks is not often dealt with. This article aims to contribute in both areas. Firstly, it offers a broad perspective on the publications of Ibero-American fact-checkers on Facebook between 2016 and 2021, focuses on the evolution of video production (n=9075) and on the views and engagement achieved by this format with respect to the rest, and relates them to changes in the News Feed algorithm. Secondly, it proposes a content analysis to identify formal and thematic elements in the most popular videos in the same period (n=414) and relates them to previous research. Our results show significant similarities in popular videos, but also changes in video production, a generalized decrease in the ratio of views and a drop in the interaction rate more accentuated than in all the publications of the period. Although the focus of this research does not allow us to make direct causal inferences, the trends identified coincide with the changes in the Facebook News Feed algorithm that were made public in those years.

Scanning or Simply Unengaged in Reading? Opportune Moments for Pushed News Notifications and Their Relationship with Smartphone Users' Choice of News-reading Modes

News notifications on smartphones provide a convenient way to stay informed, but their delivery timing can influence user engagement. Despite this, research on the impact of notification timing on reading behavior remains limited. Therefore, we developed NewsMoment, a news aggregation app that monitors user reading patterns and sends news notifications. Our experience sampling study with 46 NewsMoment users revealed four distinct reading modes: typical, comprehensive, scanning, and unengaged. Deep reading, encompassing typical and comprehensive modes, more often occurred during self-initiated browsing rather than through pushed news. Interestingly, shallow reading modes - unengaged and scanning - showed varying prevalence, associated triggers, and engagement, despite their similarities. Importantly, unengaged reading persisted regardless of users' perceived moment opportuneness, whereas scanning reading was more common during inopportune moments. These findings suggest that identifying opportune moments for news reading may primarily reduce scanning reading, without substantially impacting unengaged reading.

“Dangerous in the Minds of Barrio People”: Empowering Bicultural Latino Identity on Spanish-Language Television in Los Angeles, 1960–1990

Abstract This article focuses on the bicultural Latino identity-building discourses promoted by TV stations KMEX Channel 34 (Spanish International Network/Univision) and KVEA Channel 52 (Telemundo) through news and public affairs programming in metropolitan Los Angeles, home of the United States’s largest concentration of Latinas and Latinos. The two stations’ news and public affairs shows are historically significant for being their most extensive locally produced content and offering a valuable audio/visual representation of Latino Southern California’s cultural development from the early 1960s to late 1980s. Both stations reflected their social context, broadcasting to an overwhelmingly Mexican-heritage audience; but producers and investors at KMEX-34 and KVEA-52 also consciously sought a wider Spanish-speaking audience in Southern California by framing viewers as politically empowered bicultural U.S. Latino subjects defined by a positive affirmation of their cultural background (broadly inclusive of all of Latin America) and their integration into U.S. life through participation in electoral politics and other civic processes. Combining archival research consisting of early scripts, office memoranda, program reviews, viewer letters, oral history interviews, and analysis of extant video recordings preserved in the UCLA Film and Television Archive’s News and Public Affairs Collection, I historicize the ways locally produced programming on Spanish-language TV framed and promoted notions of an additive and empowering bicultural U.S. Latino identity.

Platform Configuration and Digital Materiality: How News Publishers Innovate Their Practices Amid Entanglements with the Evolving Technological Infrastructure of Platforms

ABSTRACT Technology companies and their platforms are important digital intermediaries for news publishers. Platforms as technological infrastructures have digital materiality and how publishers organise their innovations have become entangled with the digital materiality of platforms. This six-year longitudinal study (2015–2021) examines this entanglement by analysing how and why a news publisher innovated its practices vis-à-vis the changing digital materiality of Facebook and Instagram. Employing practice theory and the platform configuration framework, this study assesses the publisher’s configuration of its editorial activities in relation to shifts in the digital materiality of Meta’s platforms and the reflexivity of newsroom staff. Using a mix of qualitative methods, this study finds that the publisher both embraced and resisted the platforms’ algorithmic changes. While it enhanced video production and distribution to align with the platforms’ preference for audio-visual content, the publisher also developed new practices to counterbalance its reliance on Facebook for news distribution and digital advertising revenue, especially after Meta’s downranking of publishers’ content on News Feeds and the introduction of restrictions for Instant Articles. This study concludes that while publishers are circumscribed by the changes in the platforms' digital materiality, they strategically innovate specific practices to both leverage platforms and reclaim or maintain independence from them.

A STUDY OF NEWS CONSUMPTION PATTERNS AMONG COLLEGE STUDENTS

India will have 1 billion smart phone users by 2026, according to a recent Deloitte report Big bets (2022). As for news consumption, a June 2022 article in Mint says that around 72% on Indians access news through smartphones, and only 35% via computers. News aggregator platforms and apps such as Google News (53%), Daily Hunt (25%), In Shorts (19%), and News Point (17%) are also gaining currency Jha (2022). Even though the use of mobile phones is increasing, a 2022 study conducted by the Reuters Institute in 46 markets says that the trust in news is falling, especially with the younger age group. Youth Avoiding News, Says Reuters Study. (2022). The report says, “These groups are critical audiences for publishers and journalists around the world, and for the sustainability of news, but are increasingly hard to reach and may require different strategies to engage them.” This study also marks the difference in the consumption pattern between social natives (18 to 24) and digital natives (25 to 34). While digital natives still stick on to Facebook, social natives have moved to other new networks. Many news organizations are struggling to remain relevant to this group of youth. Using the quantitative survey methodology, this study aims to track news consumption patterns among youth.

News Aggregation using Web Scraping News Portals

In today's fast-paced world, staying informed through reliable and high-quality news is important. However, with an overwhelming number of online publishers and platforms, accessing information from multiple sources can be time-consuming and challenging. To simplify this process, news aggregators have emerged as a solution. These platforms allow users to customize their news consumption by selecting preferred websites and receiving curated articles from those sources in one centralized location. This saves valuable time and effort but also streamlines information gathering in our daily routines. To develop an influential online news aggregator, web scraping for data structuring is crucial. Web scraping involves analyzing a website's HTML structure to extract the desired data. By understanding the underlying structure of web pages, developers can fetch relevant information such as article titles, summaries, authors, and publication dates. People often struggle to find valuable news sources amid the vast sea of information available online. News aggregators address this issue by providing personalized news feeds based on individual interests. Users serve as valuable resources for knowledge tailored to specific preferences. While news aggregators are widely recognized for their popularity and utility, there is still room for improvement in their role in software development. Enhancements could include refining algorithms for more accurate content curation. In addition, Users could include implementing better user interfaces for enhanced user experiences and exploring innovative features like sentiment analysis and topic clustering. By continually iterating and improving upon existing news aggregator platforms, developers can ensure users have access to timely, relevant, and trustworthy news conveniently and efficiently.

Influencing recommendation algorithms to reduce the spread of unreliable news by encouraging humans to fact-check articles, in a field experiment

Society often relies on social algorithms that adapt to human behavior. Yet scientists struggle to generalize the combined behavior of mutually-adapting humans and algorithms. This scientific challenge is a governance problem when algorithms amplify human responses to falsehoods. Could attempts to influence humans have second-order effects on algorithms? Using a large-scale field experiment, I test if influencing readers to fact-check unreliable sources causes news aggregation algorithms to promote or lessen the visibility of those sources. Interventions encouraged readers to fact-check articles or fact-check and provide votes to the algorithm. Across 1104 discussions, these encouragements increased human fact-checking and reduced vote scores on average. The fact-checking condition also caused the algorithm to reduce the promotion of articles over time by as much as −25 rank positions on average, enough to remove an article from the front page. Overall, this study offers a path for the science of human-algorithm behavior by experimentally demonstrating how influencing collective human behavior can also influence algorithm behavior.

Good News – Bad News. Proportion between Positive and Negative Headlines in the Global News Feed (Based on the Google News Aggregator)

This article tests the assertion about the dominance of the negative agenda in the media space. To this end, Google News, the world’s largest news aggregator, was taken into account, on the basis of which the ratio of positive and negative news headlines in the information flow was analyzed. Three stages of research were conducted, each one week long with 16 country versions of Google News. In total, it was considered a little less than half a million news headlines. According to the results, 70.54% of the headlines turned out to be positive. This outcome questions the stereotype of the dominance of negative agendas in the global news stream. As it turned out, the ratio of positive and negative news depends on the language (versions of Google News in one language showed very similar results) and the day of the week (the news is a little more positive on Sundays), but does not depend on the level of freedom of speech, cultural and geographical factors.

Abundance and some parameters of reproduction for pink salmon <i>Oncorhynchus gorbuscha</i> on Sakhalin coast of the Tatar Strait

Annual data on biomass and number of pink salmon in catches, mainly landed from their feeding aggregations by trap nets, are presented for the Sakhalin coast of the Tatar Strait in 2003–2022. Number of the spawners entered to the rivers of this coast and number of the juveniles migrated downstream are assessed for some years. Dynamics of the pink salmon run to the rivers does not correlate with the dynamics of this species catch in the coastal waters that confirms the earlier conclusion about mixed origin of pink salmon in the feeding aggregations at the coast. The basis of pink salmon catches in recent years is the fish originated from the rivers of the adjacent mainland coast, whereas the stocks of pink salmon on the Sakhalin coast of the Tatar Strait are in depression caused by unfavorable environmental conditions of their spawning and embryonic development (as the spawning grounds erosion by extreme floods) and subsequent feeding of juveniles in the coastal waters (as heavy storms in the time of their migration from rivers).

Worldcrunch

Worldcrunch is a curated collection of international news stories from over 40 countries and 15 languages, translated into English by multilingual journalists. Based out of Paris, France, Worldcrunch was launched in 2011 and was traditionally focused on individual subscribers. However, this year Worldcrunch is prioritizing institutional site licenses. Worldcrunch is designed as a day-to-day news source rather than as a research database. Searching, printing, and downloading capabilities are limited, but Worldcrunch provides many ways to browse for articles, and an advanced search feature is forthcoming. Thankfully, no advertisements are visible to subscribers, unlike with many US-based newspapers. Pricing is reasonable for both individuals and institutions, and the insight into international news is unbeatable. This comes from the combination of news stories written by local journalists, English translation by multilingual journalists, and an editorial staff with a keen eye for stories that have a commonality to readers across all cultures.

Facebook’s News Feed Algorithm and the 2020 US Election

Facebook plays a substantial role in shaping exposure to news and information, a role that is particularly significant during democratic elections. This article aims to characterize Facebook’s contribution to news exposure during the 2020 US election. Using internet browsing data from 1.3 million desktop panelists in the United States, we analyze audience flows from Facebook to news and information publishers ( N = 3,247 websites) labeled according to quality (high or low) and geography (non-local or local). We find that Facebook ranked as the fourth largest external referral source to publishers. Our analysis further shows that, among the panel of 1.3 million US adults visiting these publishers via desktop computers, Facebook disproportionately amplified low-quality publishers and extended their audience reach during the study period. We also find evidence that Facebook’s intervention to improve “news ecosystem quality” by tweaking their News Feed algorithm was associated with fewer daily visits to low-quality publishers, but seemed to have a similar effect on other publishers and, according to Facebook-provided data, had little effect on month-over-month traffic. We discuss limitations, open questions about platform impact, and directions for future work.

Манипулятивные приемы в кликбейтных заголовках интернет-СМИ

The paper deals with manipulative means in clickbait headlines of online media. The aim of the article is to distinguish and examine the used in implicit clickbaits manipulative means which help to cultivate the reader’s interest and make them pass to the article. The relevance of the current investigation is determined by a wide spread of clickbaits in media cyberspace, the users’ negative attitude to them and their constant evolvement for concealment and formal conformation to the criteria of non-blocked headlines. The analysed material comprises the clickbait headlines in online media whose announcements are presented on large Russian news aggregators. The material was selected within 2018-2022. The overall number of investigated materials is 3050 units. The material is analysed using such methods as introspection, conceptual modelling, unselected and purposive samplings and interpretative analysis. The analysed material shows that according to the degree of connection with the article clickbat headlines can be divided into 4 types: 1) clickbait headlines of “universal” wording applicable to texts with any content; 2) clickbait headlines which are not connected with the content of the article, 3) clickbait headlines connected with the content of the article via their non-sensational parts; 4) clickbait headlines reflecting the content of the article. Three first types are considered obvious clickbaits while the last one is an implicit clickbait. The manipulative means which give an opportunity to make a clickbait implicit include manipulation with word meanings, stylistic and grammatical manipulations, provoking logical mistakes, reticence, overstatement, neglecting the time factor, taking a statement out of context, substitution of concepts, assimilation of headlines to typical associations.

Effective News Text Summarization Techniques

n proposed work, we successfully implemented a news text summarization system using Natural Language Processing (NLP) techniques and the Latent Semantic Analysis (LSA) algorithm. The purpose of our project was to extract important information from a large volume of news articles and present it in a concise and easily understandable manner. To achieve this, we utilized the LSA algorithm, which is known for its ability to capture the underlying semantic structure of text. LSA employs a mathematical model to analyse relationships between words in a document, creating a semantic representation where words with similar contexts are grouped together in a vector space. The LSA-based summarization process involved several steps. First, we pre-processed the news articles by removing stop words, punctuation, and other non-relevant elements. Then, we constructed a term-document matrix, where rows represented words and columns represented documents, with matrix values representing word frequencies. Next, we applied Singular Value Decomposition (SVD) to the term-document matrix. SVD helped reduce the matrix's dimensionality by identifying the most important latent semantic concepts. This resulted in a lower-dimensional representation that captured the essential information. Finally, we identified the most important sentences in the news articles by measuring the cosine similarity between each sentence and the summary. Sentences with the highest cosine similarity scores were selected as summary sentences. The proposed system demonstrated the effectiveness of the LSA algorithm for news text summarization. By capturing the semantic structure of the text, it generated summaries that allowed users to understand the key points of a news article quickly and easily. Our implementation had practical applications for content recommendation systems, news aggregation platforms, and personalized news feeds. However, it is important to acknowledge the limitations of the LSA algorithm. It may struggle with handling idiomatic expressions and can be sensitive to the quality of the input data. These considerations highlight the need for ongoing research and development to enhance the performance and robustness of news text summarization systems.

Algorithms patrolling content: where’s the harm?

ABSTRACT At the heart of this paper is an examination of the colloquial concept of a ‘shadow ban’. It reveals ways in which algorithms on the Facebook platform have the effect of suppressing content distribution without specifically targeting it for removal, and examines the consequential stifling of users’ speech. It reveals how the Facebook shadow ban is implemented by blocking dissemination of content in News Feed. The decision-making criteria are based on ‘behaviour’, a term that relates to activity of the page that is identifiable through patterns in the data. It’s a technique that is rooted in computer security, and raises questions about the balance between security and freedom of expression. The paper is situated in the field of responsibility of online platforms for content moderation. It studies the experience of the shadow ban on 20 UK-based Facebook Pages over the period from November 2019 to January 2021. The potential harm was evaluated using human rights standards and a comparative metric produced from Facebook Insights data. The empirical research is connected to recent legislative developments: the EU’s Digital Services Act and the UK’s Online Safety Bill. Its most salient contribution may be around ‘behaviour’ monitoring and its interpretation by legislators.