YouTube Narratives on the Al-Shifa Hospital Raid: Grasping the Debate around the Al-Shifa Hospital Attacks

Yu Hu, Yunuo Wang, Yuxiao Ye, Yuqi Zhang

Introduction

This research examines narratives that unfold on YouTube regarding the Al-Shifa Hospital raid in Gaza carried out by the Israeli Defence Forces (IDF) on the 15th of November. This operation occurred amid the continuing conflict between Israel and Gaza that intensified on the 7th of October 2023, following a surprise attack by Hamas on Israel and Israeli citizens. After that, “Israel launched a major military campaign in the Gaza Strip with the objective of destroying Hamas - which it classes as a terrorist organization - in response to an unprecedented cross-border attack on 7 October” (Gritten, 2023). According to a news report from The Guardian, “Israeli soldiers were inside Gaza’s largest hospital on Wednesday after an early morning raid that drew fierce condemnation from the head of the World Health Organization, who called it totally unacceptable” (Sinmaz & Burke, 2023). Since then, the Israeli government has aimed to eliminate Hamas, a political militant organization that governs the Gaza Strip.

According to the IDF, Hamas fighters used the hospital for military purposes. Israeli forces carried out a targeted search at Shifa, Gaza's largest hospital, to find weapons belonging to Hamas. The military claimed to have discovered weapons and terrorist infrastructure at the hospital, and they reported the killing of several terrorists. Intense fighting and explosions occurred in the vicinity of the hospital, raising concerns about the well-being of patients and those seeking refuge. According to Omar Zaqout, an emergency room employee at al-Shifa, “More than 180 dead bodies are deteriorating and are still lying in the hospital’s yard,” (Staff, 2023). The Israel Defense Forces delivered medical supplies, incubators, and food. The World Health Organization expressed worries after losing contact with the hospital staff and emphasized concerns about the safety of patients and personnel. Israel has indicated a willingness to potentially suspend operations, process intelligence, and strengthen logistical support.

However, evidence to support this claim is contended by Hamas and humanitarian organizations. “A government spokesperson in Gaza, which is controlled by Hamas, described Israel’s advance into the hospital as a war crime, a moral crime and a crime against humanity” (Financial Times, n.d.). Given that YouTube remains the most widely used social media platform, with a global community actively engaging on the platform, we aim to explore the recent Al-Shifa Hospital incident as a focal point for our investigation due to its significant relevance to this issue. This study attempts to grasp the dominant narratives surrounding the Al-Shifa Hospital raid by analyzing videos of Arabic, Hebrew and English on YouTube.

Research Questions

What are the most reinforced narratives about the Al-Shifa hospital in Gaza that was attacked by Israel on YouTube? Sub-question: What are the most reinforced narratives in the transcripts and the comments of YouTube videos covering the attack on Al-Shifa hospital?

Methodology

By analyzing both videos and comments, this research captures a more holistic view of public opinion. Videos often reflect the creator's perspective, while comments can provide insights into the broader community's reactions and thoughts. This dual approach helps in understanding both the propagated narratives and the public's reception of these narratives.

Step 1: data collection

We used YouTube Data Tools (Rieder, 2015) to collect all the results of the query “Al Shifa hospital” in English, Arabic and Hebrew each day from November 5 to November 21 (the week before the raid by the IDF on November 15 and the week after the raid.) For each language, we ranked all the videos by their view count and selected 20 most viewed videos per language. For 3 most viewed videos per language, we used YouTube Data Tools to collect all the comments. Thus, our dataset contains 20 most viewed videos and all the comments from 3 most viewed videos of the search queries in English, Arabic and Hebrew.

Step 2: data curation

For the dataset per language, we employed custom prompts on the GPT tool through Prompt Compass (https://promptcompass.digitalmethods.net/), “a tool designed to leverage Language Learning Models (LLMs) in digital research tasks. It accomplishes this by offering access to diverse LLMs, supplying a library of prompts for digital research, and enabling users to apply these prompts to a series of inputs. The tool offers two categories of LLMs: local LLMs and platform APIs. Local LLMs can be used to provide stable and reproducible results, facilitate in-depth analysis, and support a robust interpretation. [...] The library of prompts comprises prompts extracted from academic literature and other research, each of which is linked to its respective source. [...] The tool accepts user inputs either as text lines or CSV files, to be processed with the chosen LLM and prompt, with each line treated separately” (Erik Borra, 2023). Arabic content was generated using ChatGPT 3.5 Turbo, while English and Hebrew contents were produced using ChatGPT 4 due to the extensive text size, requiring a model with an expanded context window.

1. extract the transcript of each video;
2. using gpt to generate the narrative of the transcript of each video and the 300 most liked comments of 3 most viewed videos.

The exact custom prompt for the ‘narratives’ we used is as follows:

Objective: As a scholar in narratology analyzing YouTube video transcripts, your task is to extract narratives and various other entities and labels from video transcriptions. You will be presented with a video transcript which you need to process as follows.
Task Steps:
- Extract Title: Capture the video title to understand the overarching theme.
- Extract Geographical Entities: Identify continent and country names, including acronyms (e.g., "USA," "UK").
- Extract Thematic Terms: Focus on religion, ethnicity, and nationalistic terms (e.g., "Christianity," "Muslim," "nationalism").
- Extract stance: Determine whether the sentiment is pro, against, or neutral regarding a specific entity (e.g., "pro-Israel," "anti-Palestine").
- Identify the 'Other': Highlight who is portrayed as the antagonist or villain in the narrative.
- Identify the 'Hero': Highlight who is portrayed as the protagonist or hero in the narrative.
- Extract Narratives: Identify distinct narratives.
Output Format: Generate a JSON with columns "transcript", "Geographical entities", "Narrative", "Thematic terms", "Identified Other", "Identified Hero", "Stance", "Other", "Hero"
Note:
Apply critical analysis to categorize underlying themes and narratives. Do not provide explanations, only labels for stance, identified other, and identified hero. If there is no villain or hero, return an empty string for the respective fields.

Step 3: data analysis

For all the narratives of the transcript or the comment in each language, we grouped them into 3-5 “main narratives”. Raw Graph was used to visualize the main narratives of transcripts and comments.

Transcripts:

Figure 1. Narratives of the top 20 viewed Youtube videos (ENG)

Figure 2. Narratives of the top 20 viewed Youtube videos (HEB)

Comments:

Figure 3. Narratives of the top 20 viewed Youtube videos (ARB)

Figure 4. Narratives of the top 300 most-liked comments under the videos (ENG)

Figure 5. Narratives of the top 300 most-liked comments under the videos (HEB)

Figure 6. Narratives of the top 300 most-liked comments under the videos (ARB)

To compare narrative differences among languages and between the transcripts and comments for each language, we counted the sum of view counts of each “main transcript narrative” and the sum of like counts of the “main comment narrative”. To get more details about the videos and comments, we used the word cloud function of Orange Data Mining (Demšar et al., 2013) to get the frequency of meaningful words in the transcript of 20 most viewed videos and 300 most liked comments under 3 most viewed videos. Using 3-5 meaningful words with highest frequency as the seeds, we manually search the sentences in the transcript or the comment with these highly frequent words and categorize the narrative strategy that are used by the transcript or the comment per language, the commonly used words in the transcript are “hospital”, “army”, “hamas”, “Israel/Israeli” and “Gaza”.

Figure 7. Most used words in the 20 most viewed videos

Figure 8. 300 most liked comments under 3 most viewed videos

In addition, to examine what words co-occurred more in the comments of these videos, we also conducted bigram analysis of the top 300 most liked comments through Orange Data Mining (Demšar et al., 2013). After filtering bigrams of which the frequency is under 5, we summarize the top 10 bigrams of each language.

Figure 9. The top 10 bigrams of the top 300 liked Arabic comments

Figure 10. The top 10 bigrams of the top 300 liked English comments

Figure 11. The top 10 bigrams of the top 300 liked Hebrew comments

Theoretical Framework

Some scholars on the fake news have put forward the concept of “post-truth” to refer to “circumstances in which objective facts are less influential in shaping public opinion than appeals to emotion and personal belief” (Boler & Davis, 2018). There are a lot of frameworks to embody two strategies and their effects on the audience. Based on the spectrum with two poles of “truth” and “affect”, we could define the two different narrative strategies by its intention and map out the narrative strategy of the videos as well as whether or not they elicit the intended reaction among the YouTube users.

Some research also has pointed out 3 different persuasion strategies of the Youtube videos including logos (logical arguments) or pathos (emotion), and ethos (the credibility of source) (English et al., 2011), which could be discerned through the relationships between the transcript and comment of the videos. For our dataset, logos and pathos could be ascribed into the strategy of evidence and the strategy of sentiment. And the ethos is combined with these 2 strategies instead of taking effect independently.

To narrow these strategies down the attitude of sentences, we could borrow the framework of discourse analysis which describes three dimensions of attitude: affect (expressing emotion); judgment (assessing persons or behaviors) and appreciation (valuing an object or phenomenon) (Inwood & Zappavigna, 2023; Muinao & Ratnamala, 2023). These dimensions could help us to conclude the dynamics of the audience's reaction. The attitude of judgment points to the strategy of evidence and the attitude of affect points to the strategy of sentiment, though both strategies could include subjective claims of standpoint which ought to be ascribed to the attitude of appreciation.

Findings

The narrative-analysis on the transcripts showed that the main narrative for the videos about ‘Al Shifa hospital’ in the English search query was: ‘Hamas uses civilian infrastructure for military purposes’. In the Arabic query the main narrative was: ‘Israel is attacking civilians in Gaza’. For the Hebrew search query, the main narrative was: ‘Hamas uses hospitals for military infrastructure’. These were the narratives that reached the most viewers.

We also analyzed the narratives in the top 300 most liked comments under the top 3 videos per language query. This gave us an interesting insight into what narratives circulate amongst Youtube-users and to what extent they differ from the narratives that arose from the videos. Under the videos from the English search query, the narrative ‘Israel is weaponizing the memories of the holocaust’ came up most often as a narrative in comments. For the Arabic query, ‘Prayers and empathy for humanity’ came up most. Under the videos from the Hebrew query, the dominant narrative was ‘Support and advocacy for Israel and IDF’.

The results of the bigram analysis are as follows. The top 10 co-occurring words of Arabic are basically three types: religion, solidity, and protecting children. The top 10 English bigrams mainly show the doubts and objections of audiences to Israel. The result of Hebrew comments is more complex, both for-Palestine and for-Israel bigrams occur in the top 10. In addition, bigrams which mean “earth” and “planet” occur frequently in Hebrew comments.

Discussion

(1) The highlighted political stance of main narratives in different languages

According to our findings, the dominant narratives in different languages show different positions: the dominant narratives in English videos tend to oppose Palestine. Anti-Palestine narratives such as "Hamas uses civilian infrastructure for military purposes", "Hamas violence against Israeli civilians" and “Israel as victim of power groups in the middle east” featured most prominently in the English videos. The second is neutral reporting on AI-Shifa.

These videos are mostly posted by professional news channels which emphasize their objectivity and neutrality, or from channels from countries with less interests related to Palestine and Israel, such as India. And the anti-Israel narrative is the least viewed. Thus, the English videos are overall anti-Palestine, which is consistent with the position of the governments of the major English-using countries, i.e., the U.S., the U.K., Australia, etc., on the Palestinian-Israeli conflict.

Similarly, the Hebrew videos exhibit a bias against Palestine. These videos are characterized by a high degree of uniformity in terms of narrative type. Among the videos in Hebrew, "Hamas uses hospitals for military infrastructure" has received far more views than any other types of videos. Even the second most viewed narrative, "Hamas uses hospitals as human shields", has more views than all non-anti-Palestinian videos combined. The Hebrew videos are highly anti-Palestinian, probably due to the need for Israeli war mobilization. The dominant narrative in Arabic videos lean towards being critical of Israel. In contrast to Israel, the most viewed narrative in Arabic videos is "Israel is attacking civilians in Gaza". At the same time, Arabic-language channels countered the anti-Palestinian narrative in English and Hebrew by posting a large number of videos arguing that "Israel is spreading fake news". In addition, videos about solidarity and victory have received a high number of views. The positions of these narratives are basically aligned with the interests of the Arab countries.

The stances of the dominant narratives in all three languages are similar to their countries. The perspectives presented in the dominant narratives of Arabic and Hebrew comments, as well as their corresponding video contents, exhibit a fundamental alignment. What is interesting is that the comments in English diverge considerably from the video contents. The top 2 main narratives of English comments are anti-Israel and pro-Palestine. Viewers of the English videos criticized Israel for its "weaponization of its holocaust memories" as well as expressing support for Palestine through slogans and prayers. This disagreement may be due to various reasons. First, as English is a universal language, its speakers are not necessarily native speakers, so many of the English comments may not be from English-speaking countries. Second, the decline of media credibility is also one of the possible reasons for the divergence between video content and comments. The credibility of the media declines when users realize they have been exposed to fake news (Tandoc Jr. et al., 2021). According to the fourth-ranked narrative, "CNN reporter is fired for advocating for Palestine," English-speaking users have already had doubts about the credibility of news outlets such as CNN. As a result, they may no longer trust these outlets and may instead resist and oppose their advocacy. Third, astroturfing is a potential cause as well. It refers to fake product or political reviews posted by paid writers on social media for commercial or political interests (Peng et al., 2017). The difference between the content and comments of English videos is possibly caused by some fake and paid comments.

(2) The strategy of evidence and the reaction of users

Among the transcript of the 20 most viewed videos, we could observe both narrative strategies in English query. One is the “strategy of evidence”: to quote the practical evidence offered by the IDF to advocate the narrative “Hamas uses civilian infrastructure for military purposes’. The main argument includes the statement of fact: “there is a tunnel under the hospital” or “there was a Hamas armory”. This strategy also takes up the main trend in the transcript of videos under the Hebrew query, while the difference is that the Hebrew transcript tends to directly and completely showcase the pro-Israel argument and justification that are officially released by the IDF instead of quoting the clips of these content and re-contextualizing them in a more neutral tone.

Although the source of content is similar, the Hebrew comment under the videos of the Hebrew query is mostly expressing support for Israel, for instance, “General Daniel, I support you all the time...I stand with you forever”. Though there are also a few comments in other languages that express criticism or doubt against the Israeli propaganda which has made up of the third most dominant narrative “support for Palestine and criticism of Israel''. According to the bigram of the comments under the Hebrew videos, the comments are mainly the phrases of “appreciation” in different languages and standpoints without typical expressions of “affect” or “judgment”.

There are more conversations and debates going on in the English comments, making the discourse more controversial and diversified. Taking “Holocaust” as an instance, it comp up as the high-frequency word because the criticism of “Holocaust denial” is used by the pro-Israel comments and the very sensitive criticism of “Holocaust card” is used by the pro-Palenstine comment.

Moreover, the English comment discusses more on the significance of the truth and the value of the news, including the doubt or the trust for the information source, which means that the dominance of the logos of the strategy of evidence takes over the ethos (English et al., 2011). This corresponds to the third most dominant narrative of “CNN reporter is fired for advocating for Palestine”. Though the transcripts of the videos endeavor to take a neutral tone, the comments under them are dominated by the reactive doubt and dissatisfaction against the opposite viewpoints. Aside from the “holocaust”-related phrases, the bigram of the comments under English videos is also mainly showing the phrases in the fact statements, which is a combined attitude of “judgment” and “appreciation” (Inwood & Zappavigna, 2023), which shows the dominant effect of the strategy of evidence.

In the meantime, the dominant narrative in the Arabic videos whose strategy could also be ascribed to the strategy of evidence is “Israel is spreading fake news” which is achieved by quoting the claims of Israel and contradicting it with their further fact-checking. There is only one video in the 20 videos of English queries that has the same narrative. Apart from these fact-checking results, the Arabic query also outputs videos with the uncanny conspiracy which is for the narrative “Israel is attacking the civilians in Gaza” — “75 years of occupation of the Land and genocide of the Palestinian People. More than 4000 children killed, more than 10000 peaceful civilians killed.”

(3) The strategy of sentiment and the reaction of the users

Another narrative strategy of the videos under English query is the strategy of sentiment: to claim that Hamas is using the hospital to execute the Israeli hostages, which is not only the narrative “Hamas violence against the Israeli civilians” and “Israel as victim of power/groups in the Middle East”, but still radically points to the most prevalent narrative “Hamas uses civilian infrastructure for military purposes”. For instance, “Hamas murdered Noa (an Israeli hostage) in the hospital...IDF will do everything to bring all the hostages back home”. The style of comments are also different under these two types of videos since the comments under the latter type of videos are dominated by the sorrow or sympathy for the victims. While these sentimental expressions also incline to political viewpoints when they extend to their support or wish for IDF or Israel’s military actions. The strategy of sentiment is seldom used in the results of Hebrew query, but there is also a single video focusing on the hostage which also elicits the sorrowful reactions in the comment. As for the narrative “neutral finding on Al-shifa: fact finding” that takes a large part of the videos under English query, the main strategy is to demonstrate the severe casualty or suffering of the civilians in Gaza, many with sorrow and sympathy, which is also the strategy of sentiment. The similar expression similarly exists in the videos under Arabic query, some even use the same video (For instance, the video of the English nurse crying while reading the messages from Al-shifa hospital). The style of the comment under these videos are also similar in English and Arabic with the users of both queries expressing their prayer, sympathy or wish for the victims and their hate against the war. With a large proportion of these comments, the most dominant narrative in comments under Arabic videos is “prayers and empathy for humanity”. The result of bigram has strongly shown that the comments under Arabic videos are highly dominated by the calling and prayer within faith contexts - a combination of “emotion” and “appreciation” (Inwood & Zappavigna, 2023), which suggests the dominant effect of the strategy of sentiment.

What is noteworthy is that some comments are not purely neutral and humanistic sentiments but combined with the claims with political stances such as “I do not think that God is unaware of what the wrongdoers do. He only delays them for a day when eyes will be clear”, which is reflected in the second and the fourth most dominant narrative “support for Palestine & Humanitarian Concerns” and “Criticism for Israel’s war crimes”.

Conclusion

According to our findings and discussion, for the mainstream media, though there is a difference in political stance and tendency among the queries in different languages, the narrative strategies which include fact-checking and sentiment-expressing could be shared. Though usually trying to keep a neutral and objective attitude, the motivations and biases are suggested by the selection of narrative since almost every narrative has the function of justifying the stance of pro-Israel or pro-Palesitine. The relationship between the narrative and its implicit political orientation determines the narrative strategy it is ascribed.

As the concept of “post-truth” has indicated, the strategy of sentiment elicits more consensus in the comments than the strategy of evidence and the certain political tendency which the sentiment-expressing narrative underpins would also stand out in the users’ reaction. However, though the strategy of evidence tends to elicit more drastic discussion on the controversial issue, the extent to which the issue is raising conflict among the audience varies in different contexts, since the similar narrative raises more affective resonance and consensus in Hebrew comments than English comments.

For the users of YouTube, the political arguments and sentimental expressions are always interlocking. The political argument is prompted by the expression of sentiment and the sentiment-expression could also be transformed into the political tendencies. In the discussion of controversial issues, the incorporation of people’s emotional expression like doubt and rage may further obfuscate the truth and their justification. While, the sentiment of praise or sorrow may raise more resonance and lead to the consensus in opinion. Therefore, for further research, the effect and function of different types of sentiment could also be an interesting direction.

-- Main.BernRieder - 19 Feb 2024