Leveraging Artificial Intelligence in Social Media Analysis: Enhancing Public Communication Through Data Science

Abstract

This study examines the role of AI tools in improving public communication via social media analysis. It reviews five of the top platforms—Google Cloud Natural Language, IBM Watson NLU, Hootsuite Insights, Talkwalker Analytics, and Sprout Social—to determine their accuracy in detecting sentiment, predicting trends, optimally timing content, and enhancing messaging engagement. Adopting a structured model approach and Partial Least Squares Structural Equation Modeling (PLS-SEM) via SMART PLS, this research uses 500 influencer posts from five Arab countries. The results demonstrate the impactful relationships between AI tool functions and communication outcomes: the utilization of text analysis tools significantly improved public engagement (β = 0.62, p = 0.001), trend forecasting tools improved strategic planning decisions (β = 0.74, p < 0.001), and timing optimization tools enhanced message efficacy (β = 0.59, p = 0.004). Beyond the technical dimensions, the study addresses urgent ethical considerations by outlining a five-principle ethical governance model that encourages transparency, fairness, privacy, human oversee of technologies, and institutional accountability considering data bias, algorithmic opacity, and over-reliance on automated solutions. The research adds a multidimensional framework for propelling AI into digital public communication in culturally sensitive and linguistically diverse environments and provides a blueprint for improving AI integration.

1. Introduction

Public communication is progressively formed by social media platforms, establishing a critical need for advanced analytical tools that can interpret user attitude and guide strategic content. Nevertheless, most AI methods struggle to accurately analyze Arabic-language content due to language complexity and cultural nuances, raising concerns about ethical transparency, reliability, and bias. Artificial intelligence (AI)-driven social media analysis—using platforms such as Google NLP, IBM Watson, and Hootsuite Insights—offers the possibility of scalable insights, but their performance and ethical implications in non-Western contexts remain underexplored.

Scholarly arguments in the 2020–2025 period have already foreshadowed these changes, even when the full implications were not yet clear. Journalism and media educators recognized the domineering need to teach students how to engage with emerging digital tools, as strong majorities agreed that media graduates must be familiar with the latest technologies and that faculty should be able to teach them how to use such tools (Babacan et al., 2025; Every Learner Everywhere, 2024; Mitchell, 2024; Wenger & Hossain, 2024). Accordingly, media researchers started mapping the changing media landscape: Deuze (2003) recognized distinct methods of online journalism with new publishing dynamics, while Pavlik (2004) defined a “sea change” in journalism as media convergence blurred the lines between producers, audiences, and content. These early studies highlighted gaps in understanding how such new technologies like AI would affect journalistic practice, ethical values, and education outcome (Warschauer, 2003). In MENA contexts, academics noted that nascent digital media were starting to reshape public discourse and community undercurrents, while formal journalism training in the region lagged behind regarding these innovations (Eickelman & Anderson, 2003).

Although several studies (Gerlich et al., 2023; Deiner et al., 2023) have explored AI in digital communication, few have focused on its application to Arabic-language content. Furthermore, the ethical and operational challenges of using AI tools for public communication in non-Western contexts such as the Arab region remain underexplored (Alhindi et al., 2023).

Accordingly, with the adoption of AI for real-time examining and strategic messaging in social media, understanding its performance in linguistically complex regions like the Arab region is needed. The fast digitalization of public organizations and the rise of Arabic content demand tools and applications that are not only technically competent but are also ethically approved.

This study is motivated by the urgent need to critically assess the performance and ethical considerations of AI-based social media analysis applications and tools when applied to Arabic-language content. As Arab media outlets and influencers in the Arab region increasingly depend on AI to guide public communication strategies, the absence of transparency and accountability in these tools imposes a detailed investigation. Adopting this knowledge gap is critical in developing accountable and context-sensitive AI systems that uphold ethical values and resonate with various audiences.

In response, this study conducts a comparative evaluation of widely used AI tools such as Google NLP, IBM Watson, Talkwalker, and Sprout Social, converging on their volume for accurate sentimental detection, trend identification, and strategic development in Arabic-language social media contexts. The study not only emphasizes technical limitations but also investigates the ethical risks accompanying their deployment. Furthermore, it proposes a five-principal governance framework aimed at ensuring the ethical use of AI in multilingual media settings. This study addresses this gap by evaluating the actual application of AI-powered analytical tools in Arabic social media environments, with a focus on their effectiveness in enhancing communication strategies and the ethical risks associated with their deployment. By bridging the domains of AI innovation and public communication, the research aligns with the urgent need to assess not only what these technologies can do, but also what they should do in the service of ethical and inclusive public discourse.

Specifically, the study pursues the following objectives: to analyze how AI tools can improve public communication strategies through social media; to evaluate the ethical and technical challenges of AI usage in social data analysis; and to propose ethical and efficient guidelines for the implementation of AI tools in public communication.

Accordingly, this study seeks to answer the following research questions:

RQ1: How do AI tools enhance the accuracy and effectiveness of public communication strategies on social media platforms?

RQ2: What are the key ethical and technical challenges associated with the use of AI in analyzing social media data for public communication?

RQ3: How can regulatory frameworks and transparent practices improve the ethical implementation of AI tools in digital public communication?

The objectives are examined through the lens of three interrelated hypotheses that underpin the study’s analytical framework. First, it is posited that the deployment of AI tools significantly enhances the precision and effectiveness of social media analytics, thereby contributing to more impactful public communication outcomes (H1). Second, the study contends that while AI offers substantial potential, its implementation is impeded by technical and ethical challenges that may diminish its operational efficacy and hinder broader societal acceptance (H2). Third, it is hypothesized that these ethical complexities can be effectively addressed through the adoption of regulatory mechanisms and transparent operational practices, thus fostering a more accountable and sustainable integration of AI into public communication strategies (H3).

Previous studies (e.g., Ahmad et al., 2023; Alawneh et al., 2023; Gerlich et al., 2023) have revealed sentiment detection powered by AI’s transformative effect on public communication, particularly within the framework of real-time engagement analytics and predictive modeling. Although AI technology can facilitate more targeted and responsive messaging efforts in this context, the performance of AI when dealing with linguistically difficult and ethically sensitive contexts—such as Arabic-language content—has also been called into question by a few studies (e.g., Alhindi et al., 2023; Deiner et al., 2023). These insights are the foundation of the research assumptions of the present investigation.

These objectives underpin three interrelated hypotheses about AI tools: they boost the precision and impact of social media analytics and public communication (H1); they simultaneously encounter technical and ethical limitations affecting their performance and acceptance (H2); and these challenges can be alleviated via transparent governance and regulatory oversight promoting responsible AI implementation (H3).

The study contributes directly to the objectives of the Special Issue by examining the intersection of media narratives, technological innovation, and the societal impacts of AI applications. By foregrounding the challenges and ethical considerations of AI in Arabic-speaking contexts, the study enriches discourse on comprehensive media practices in the digital age.

2. Literature Review

2.1. Role of AI in Social Media Analysis

Artificial intelligence (AI) is at the core of social media analysis today, offering new tools for interpreting massive amounts of user-generated data. Social media platforms generate gargantuan datasets daily, making old-fashioned manual analysis futile.

With AI in social media analysis, public communication has changed through real-time sentiment analysis, predictive analytics, and personalized engagement (Saheb et al., 2024).

AI-driven social media analytics leverages a range of modern tools and techniques to extract actionable intelligence from online conversations. Machine learning algorithms, particularly deep learning models, are routinely used in trend prediction and audience sentiment analysis (Gerlich et al., 2023). NLP enables AI systems to classify text, detect emotions, and to identify emerging conversations within large datasets (Deiner et al., 2023). SNA tools use AI to map relationships and identify influencers in online groups, helping organizations to optimize their outreach initiatives (Krajčovič, 2024). AI-oriented visual recognition programs can digest both photos and videos uploaded to social platforms, figuring out the key themes and trends within them (Anshu & Sharma, 2024).

2.2. Applications in Public Communication

According to previous studies (Richardson et al., 2025; Gerlich et al., 2023; Alhindi et al., 2023), the combination of AI and social media has brought numerous advantages to companies, such as enhanced audience analysis and content optimization. On an even more positive note, AI has also helped spread information like never before, so keeping a balanced approach where you consider the use of technology and ethical risk at the same time is paramount.

Richardson et al. (2025) have discussed that AI is significant in optimizing public communication by analyzing audience sentiment in an extensive level of detail, automatically monitoring engagement and measuring content performance. For content, automated recommendation engines optimize messaging by showing the users what they are interested in, extending the audience reach and engagement. Specifically, AI-enabled crisis communication software allows institutions to monitor negative trends at an early stage and to react well enough to minimize reputational threats (Maldonado-Canca et al., 2024). By tailoring responses and offering instant feedback, sentiment analysis and chatbots thus facilitate public involvement in enhanced digital communication strategies (Gerlich et al., 2023).

Social media data mining has a powerful impact on public communication campaigns since it provides instantaneous feedback on the reactions of audience, the effectiveness of the campaign, and social issues (Deiner et al., 2023). AI-supported analytical software tracks engagement levels and sentiment shifts, allowing policymakers and businesses to adjust their messaging strategies, according to Krajčovič (2024). Furthermore, social media data facilitates organizations in gauging the effectiveness of their communication campaigns and framing future campaigns based on predictive models anticipating audience response (Anshu & Sharma, 2024). Moreover, intelligence from AI allows any organization to conduct more evidence-based decisions and to design more focused and influential communication campaigns (Maldonado-Canca et al., 2024).

2.3. Ethical Risks and Governance

To achieve the full potential of AI, important issues need to be addressed, including data privacy, algorithmic bias, and technical constraints (Alawneh et al., 2023).

Organizations need to learn the best practices in data utilization and ethical AI adoption to maximize AI capabilities in social media analytics. According to studies by Alharbi et al. (2024), which focus on the need of investment in AI literacy and training, allowing communication professionals to interpret and apply AI-generated insights more efficiently, coupling AI tools with human expertise promotes a balanced approach to using AI that supports better decision-making and reduces bias in automated analytics. On the other hand, Gerlich et al. (2023) also suggested that the application of transparent AI models and ensuring compliance with data privacy laws will enhance public trust and credibility in AI-facilitated communication.

Despite the advantages, AI-driven social media analytics face some challenges; as pointed out by Taha et al. (2024b), ethical concerns and data privacy are two of the main challenges today. AI algorithms are based on enormous amounts of personal data, which raises concerns about permission, transparency, and other possible misuse. Algorithmic bias is another major challenge. Gerlich et al. (2023) argued that if AI algorithms are trained on biased data, they may produce biased analyses, leading to erroneous conclusions and reinforcing social inequalities.

2.4. Technical Challenges

Deiner et al. (2023) stated that technical challenges in using AI analysis for social media exist, such as processing large volumes of unstructured data from various sources or platforms. He also contended that making sure AI models recognize nuances like sarcasm, slang, and multilingual input accurately remains an open research problem in this area. However, Anshu and Sharma (2024) mention how important aspects of AI-driven social media insights would be integrated into the developmental decisions, which require qualified specialists, and many media organizations have been facing shortages of adequate AI-qualified people. Kazmi (2025) claims that the challenges of integrating artificial intelligence into social media are compounded by insufficient infrastructure and resistance to change. We require multidimensional strategies to overcome these issues, including fostering a generational culture that embraces advancement and innovation, investing in workforce training, and upgrading technological infrastructure.

2.5. Tools and Techniques Used in AI-Driven Social Media Analytics

Researchers (Gerlich et al., 2023; Deiner et al., 2023; Maghsoudi et al., 2024) highlight the diverse applications of artificial intelligence-driven analyses of social media. For example, natural language processing (NLP) and sentiment analysis are instrumental in identifying and categorizing public sentiment with precision. AI-enabled systems enable the classification of social media content into three main themes, namely positive, negative, or neutral sentiments, thus enabling generations to systematically assess public responses to events, policy decisions, and strategic campaigns. Additionally, AI-driven prognostic analytic and trend detection mechanisms authorize both businesses and policymakers to proactively identify and respond to emerging accounts and potential crises before they intensify.

AI-driven social network analysis (SNA) enables the identification of influences and patterns of information diffusion within online communities. This can be utilized by organizations to optimize their communication efficacy by engaging with key opinion leaders and maximizing message diffusion (Krajčovič, 2024). AI-driven content personalization also enhances audience engagement using messages that are tailored based on audience interests and behaviors (Anshu & Sharma, 2024).

Conversational AI and chatbots also support the facilitation of public communication by automating interaction with customers and enabling instant responses to inquiries (Maldonado-Canca et al., 2024). AI-driven bots are an integral focus of public sector organizations and companies providing fast customer service and instant communication (Hayes et al., 2021).

2.6. Best Practices for Ethical and Effective Implementation

Organizations need to adopt the best practices in data utilization and ethical AI adoption to maximize AI capabilities in social media analytics. According to studies by Alharbi et al. (2024) focusing on the investment needs in AI literacy and training, coupling AI tools with human expertise promotes a balanced approach to using AI that supports better decision-making and reduces bias. Gerlich et al. (2023) also suggest applying transparent AI models and ensuring compliance with data privacy laws.

3. Materials and Methods

The current study proceeds from a theory-driven and structured modeling perspective that seeks to assess the (operational and normative) performance of vital social media analytical tools, employed for administrative performance assessments using AI. Grounded in real-world digital communication practices, the study does not use direct field data collection methods but rests upon a repertoire of benchmarking patterns taken from recent academic and industry publications, such as Gerlich et al. (2023) and Maldonado-Canca et al. (2024), whose in-depth user interaction metrics are available, as well as twenty-first century-specific features (Gerlich et al., 2023; Maldonado-Canca et al., 2024; Kazmi, 2025).

The tools examined include Google Cloud Natural Language, IBM Watson Floristic Natural Language Understanding, Hootsuite Insights (Brandwatch), Talkwalker Analytics, and Sprout Social. The evaluation centers on the tools in certain processes, such as sentiment detection, trend foresight, strategic content placement, and strategic engagement assessment for real digital arenas.

3.1. Research Design

The study takes a theory-driven approach in the research design, with Partial Least Squares Structural Equation Modeling (PLS-SEM) for social media content analysis, using SMART PLS for this purpose. Instead of conventional empirical approaches such as surveys or interviews, it uses a structured modeling strategy and analysis of tools for comparison. The researchers expressed the importance of this area of continuation in evaluating AI tools based on categories, such as text analysis, trend analysis, timing optimization, and its design, relating to public communication outcomes including engagement, strategic planning, and message effectiveness. By feeding information to SMART PLS, the dataset has compiled and authenticated raw social media posts of public social media influencers, model comparisons, and complexity in model testing resulting from latent variables. Using such a methodology allows for ethical standards regarding data privacy and transparency to be upheld in analyzing digital communication practices.

3.2. Data Processing Procedures

Before analysis, the dataset was pre-processed to remove emojis, hashtags, URLs, and duplicate posts. Posts were normalized to MSA where possible using light stemming. Each post was then fed into the corresponding AI tool’s API interface using standardized batch input formats. The raw sentiment and engagement outputs were exported and cleaned for inconsistencies. The data was then imported into SMART PLS 4, where latent constructs (e.g., public engagement, communication planning) were operationalized and analyzed using PLS-SEM with bootstrapping (5000 samples) to test path significance. Model validity was evaluated using Cronbach’s Alpha, composite reliability, AVE, and discriminant validity metrics.

3.3. Sample Design

The dataset consists of 500 influencer posts collected from five Arab countries (UAE, Saudi Arabia, Egypt, Morocco, and Lebanon) across several topics, including governance, health, education, fashion, and technology. Influencers were chosen based on validated metrics like the strength of their digital footprint, their engagement in public forums, and the breadth of the themes they covered.

The selection of 500 posts was based on prior work analyzing AI benchmarking in multilingual sentiment analysis and recommend datasets that include a trade-off between statistical relevance and manual interpretability (Gerlich et al., 2023). Influencers were chosen based on well-documented levels of activity, thematic heterogeneity, and their digital trace characterized by followers and engagement levels. This sample size gives us a very dense though not overly large corpus for testing tools and comparative experiments, as shown in Figure 1.

This allows for a breadth of analyses regarding how AI tools are being applied in regional public communication practices, demonstrating a representative cross-section of styles and dynamics of engagement.

Inclusion and Exclusion Criteria:

The criteria for inclusion were based on publicly available social media posts published by verified influencers in five Arab nations. Messages had to be in Modern Standard Arabic (MSA) for polarity rating by tools to be comparable. Furthermore, selected posts were required to include enough meaningful textual content related to public communication (e.g., governance, health, and education).

The exclusion criteria included non-textual posts (e.g., images or videos without captions), content written in dialect or mixed-language formats, and posts lacking engagement metadata (likes, shares, and comments). Posts marked private, restricted to followers, or those with unverifiable authorship were also excluded to maintain transparency and replicability.

3.4. Data Strategy Justification

A theory-based analytical design is adopted in this study, which integrates structured modeling with the empirical findings of behavioral mechanisms, tool performance benchmarks, and social media engagement metrics, as described in recent academic and industry research. Instead of direct fieldwork or primary data collection, the study is built on actual content produced by social media influencers from five Arab states. These posts were publicly available and included based on validated metrics of engagement.

The approach ensures reliability by leveraging established AI tools to process naturally occurring digital communication data while maintaining ethical integrity using open access, non-intrusive content. The dataset reflects regional diversity and thematic relevance, encompassing governance-, education-, health-, and technology-related posts. This strategy enables the study to produce insights that are both methodologically sound and contextually grounded.

To enhance transparency and provide an applied example of tool behavior, a sample application of Google Cloud Natural Language API on Arabic-language posts is included in showcasing sentiment scores and engagement correlations in real influencer content.

AI Tools Selection and Overview

This research utilized five sophisticated AI tools—Google Cloud Natural Language, IBM Watson NLU, Hootsuite Insights, Talkwalker Analytics, and Sprout Social—selected for their proven efficacy in sentiment analysis, trend identification, multimedia interaction, and content scheduling. These tools were chosen for their pertinence to Arabic-language material and their demonstrated effectiveness in public communication settings. Each tool provides a distinct analytical perspective that jointly facilitates a thorough assessment of digital audience behavior and message efficacy.

3.5. Tool Selection and Comparative Evaluation

This study selected five AI tools—Google Cloud Natural Language, IBM Watson NLU, Hootsuite Insights, Talkwalker Analytics, and Sprout Social—based on their functional alignment with the study’s objectives in enhancing digital public communication, particularly across Arabic content and diverse social media platforms.

Why These Tools Were Selected:

• Google NLP and IBM Watson NLU were chosen for their advanced capabilities in semantic analysis and sentiment detection, with strong support for Modern Standard Arabic (MSA). These tools ensure context-sensitive and culturally aware language processing, which is essential for the region-focused nature of the study.
• Hootsuite Insights and Talkwalker Analytics were selected for their real-time trend detection, cross-platform performance tracking, and multimedia content analysis, which support comprehensive evaluations of public engagement and message dissemination.
• Sprout Social was included for its strengths in historical engagement tracking, scheduling analytics, and audience interaction metrics, adding strategic insights into long-term communication impact.

Comparison with Alternatives:

These tools were benchmarked against industry alternatives like Brandwatch and BuzzSumo, which were ultimately excluded. While those tools offer basic analytics, they lack adequate Arabic language processing, offer limited or outdated API access, and do not support real-time multimedia tracking or content scheduling—all features essential for this study’s mixed-method analysis.

In sum, the selected tools provide a more robust, multilingual, real-time, and contextually aware foundation for analyzing the evolving landscape of public digital communication, especially within Arabic-speaking contexts.

The Multilingual Perspective and Arabic Language Sensitivity

A large part of the data was Arabic social media content, which explains the emphasis on multilingual capabilities of the AI tools chosen for this task. Google Cloud Natural Language API and IBM Watson Natural Language Understanding provide partial support for Modern Standard Arabic (MSA), which includes sentiment analysis, named entity recognition, and syntactic parsing.

Table 1. Functional mapping and justification of selected AI tools.

Tool	Key Function	Justification
Google NLP	Sentiment and syntax analysis	Supports MSA; strong API for entity recognition
IBM Watson NLU	Semantic mapping	Tracks conceptual frames; interprets value-laden language
Hootsuite Insights	Trend tracking	Real-time data; ideal for audience behavior mapping
Talkwalker Analytics	Multimedia interaction analysis	Cross-platform media analysis; tracks CTR and attention span
Sprout Social	Timing optimization	Scheduling precision; historical engagement modeling

presents the functional mapping and justification of the selected AI tools based on their multilingual processing capabilities and alignment with the study’s objectives.

However, even with this partial backing, most of these AI agents struggle to process Arabic dialects, including Egyptian, Moroccan, Tunisian, and Levantine. As shown by Alhindi et al. (2023), emotion and meaning in dialects and slangs are often missed if processing tools are developed on and for MSA. To help solve this problem, we only concentrated on posts in MSA, especially those which appeared in formal and official content, to increase sentiment analysis accuracy and reliability. This approach allows us to keep the analysis linguistically invariant, reflecting representative tendencies in Arabic public discourse.

While these tools offer acceptable levels of accuracy in analyzing formal Arabic, particularly in public awareness campaigns, governmental updates, and institutional communications, they still face limitations when dealing with informal dialects, colloquial expressions, and sarcasm, which are commonly present in Arabic digital discourse.

Despite these constraints, the tools were selected based on their demonstrated effectiveness in processing structured Arabic content and their compatibility with model customization and language expansion. Their ongoing updates and support for multilingual processing make them suitable for research focused on formal influence-generated content related to governance, health, and education.

The complementarity of the Arabic-language posts represents an essential depth in regional and linguistic terms, filling a well-documented gap within social media analytics research, which is traditionally biased towards English-language content. Researchers can take this framework as a basis to develop future research with more subtly tuned NLP models for different Arab dialects and extend the semantic accuracy and cultural sensitivity.

3.6. Research Model

This study is grounded in a structured research model designed to assess the influence of three core categories of AI tools on key dimensions of public communication. The model is operationalized using Partial Least Squares Structural Equation Modeling (PLS-SEM) via SMART PLS, a method well-suited for theory testing and modeling relationships between latent variables in complex datasets.

The model investigates how different types of AI tools—focused on textual analysis, trend forecasting, and timing optimization—impact public engagement, strategic planning, and overall communication effectiveness. The analysis is based on real-world data collected from social media posts by influencers in Arabic-speaking countries, allowing for the evaluation of AI tool performance within authentic digital communication contexts.

Independent variables (category of AI tools):

• NLP tools: Google Cloud Natural Language, IBM Watson NLU.
• Trend analysis tools: Hootsuite Insights and Talkwalker Analytics.
• Tools for timing optimization: Sprout Social.

Outcome variables (communication outcomes):

• Public engagement.
• Communication planning.
• Communication effectiveness.

Modeling Relationships and Hypotheses:

• H1: Text Analysis Tools- Public Engagement
• H2: Strategic Communication Planning → Trend Analysis Tools
• H3: Timing Optimization Tools → Communication Effectiveness

This conceptual framework captures the objective of the study to measure the differential effect of AI tools on public communication strategies by comparing actual social media data with insights derived from the tools to validate the posited hypotheses.

4. Results

4.1. AI Tool Observations and Interpretations Summary

This overview aims to shed light on the significant results reflected in the interactions of a specific AI tool with the real-life dataset of 500 social media posts. The core functions of each tool were used to process the dataset for its sentiment, meaning, engagement, and multimedia elements, allowing for deeper analyses of Twitter interactions. The findings provide answers to the question of how public communications are affected by the introduction of AI technologies across the digital media landscape.

As shown in

Table 2. AI tool observations and interpretations summary.

AI Tool	Field of Analysis	Key Observations (Data-Driven)	Interpretation	Related Hypothesis/Question
Google Cloud Natural Language	Sentiment detection, entity recognition, topic classification	78% of posts were positive in sentiment, and positive posts had 45% more engagement on average than neutral/negative posts.	Emotionally meaningful content motivates greater audience participation, consistent with the results of Saheb et al. (2024).	H1, Q1
IBM Watson NLU	Semantic context and conceptual mapping	Posts stressing terms like “trust” and “security” were shared 40–45% more.	By reframing the conversation in relation to its value, you establish credibility and message integrity that pushes insights like observed with Deiner et al. (2023).	H1, H2, Q2
Hootsuite Insights	Tracking trends in real-time, analytics that time	Engagement during this window peaked about 24 h after publication and drove a 50–60% increase in interaction.	Posting at the right time makes sense, considering the responsiveness of most audiences and predictive release strategies like the form in which they indicate to manage timing (Maldonado-Canca et al., 2024).	H1, Q2
Talkwalker Analytics	Leverage cross-platform insights with multimedia engagement	They achieved up to 39% longer attention spans and 15–18% higher CTR with posts combining video with external links.	These findings align with Krajčovič (2024) and Anshu and Sharma (2024), who assert that rich media formats lead to increased audience retention and actions taken.	H1, H2, Q1, Q3
Sprout Social	Content scheduling and engagement timing data	Publishing midweek, between the hours of 5–8 PM saw up to 55% higher interaction rates.	Automated scheduling tools enhance reach optimized with specific patterns of audience activity, as evidenced by Kazmi (2025) and also by Alharbi et al. (2024).	H1, H3, Q1, Q3

, The data insights explain how each individual AI tool improves a facet of public communication performance in real terms. The effects found conform with both previously reported patterns seen in the literature and in the predictable capabilities of plain language tools.

4.2. Inductive Analysis Through AI Tool Outcomes and Reference Works

Through the use of the five AI tools, the patterns of how various analytical functions facilitate public communication really start to become clear. The strengths of each of the tools reflected different aspects of audience engagement, content engagement, and message delivery, and the results align with the literature.

When we applied Google Cloud Natural Language to detect the emotional tone across influencer posts, we found that it was highly effective, detecting a large majority of content as having a positive sentiment which connects to higher levels of engagement. This supports the concept mentioned by (Saheb et al., 2024) and Gerlich et al. (2023), with this further confirming the predictive power of sentiment-based targeting.

IBM Watson NLU infused semantic depth by pinpointing value-laden language—especially mentions of “trust” and “security”—that highly predicted whether content would be shared. These results indicate that how news organizations frame value arguments can strongly influence how the public perceives them and their institutions (which echoes the findings of Deiner et al. (2023) and Richardson et al. (2025)). These findings highlight how congruence between three different components of how a message is framed, and of how a message is expected to be framed, builds message integrity.

Both Hootsuite Insights and Sprout Social mentioned the temporal dynamics of audience interaction with the platform, concerning information like when you should post your articles—the best days were mid-week, and posting in the evenings achieved at least 30% more engagement. This corresponds with work already carried out by others (e.g., Maldonado-Canca et al., 2024; Kazmi, 2025) on the importance of timing to optimize expressiveness by maximizing predictive scheduling for communicative effect.

Talkwalker Analytics also provided insight into the effects of media richness. It is not the first time that we have analyzed videos together with external links; posts show that they captured viewers’ attention for longer durations and were clicked more than regular links, recording the communicative effects of video. Such findings are in alignment with the sentiments expressed by Krajčovič (2024) and Anshu and Sharma (2024), who promote the introduction of all content types in a holistic approach to strengthening audience connection.

Collectively, these findings offer strong empirical support for the study hypotheses (H1, H2, H3) and inform the core research questions (Q1, Q2, Q3). Their most impactful insight is on how different AI tools, used in an intentional way, can shape a holistic, evidence-based public communication approach that is sensitive to the audience and aware of the context.

4.2.1. Model-Based Conclusion and Link to SMART PLS Structural Analysis

The patterns revealed in this tool-by-tool interpretive analysis are augmented by the SMART PLS structural model, as depicted in Figure 2, which quantifies the effect of each tool group on communication outcomes.

SMART PLS Structural Model:

Figure 2. Structural model (SMART PLS).

Figure 2. Structural model (SMART PLS).

4.2.2. Interpretive Commentary

The results of the model confirm the following:

• Generally, the text analysis tools (Google NLP and IBM Watson NLU) have a large effect on public engagement by confirming that the sentiment and semantic analysis play a prominent role in formulating digital interaction.
• Trend analysis tools including Hootsuite Insights and Talkwalker Analytics can also facilitate strategy by homing in on real-time audience behavior and media performance data.
• Communications ops tools (Sprout Social) foster connection by finding the perfect timing to distribute your content.

Statistically supported relationships substantiate interpretative insights confirming the three hypotheses (H1, H2, H3) of the study. The alignment between these two separate approaches forms a missing part of the argument: it evidences that the application of the new powerful AI instruments into the communication process, especially with the aim of projecting success metrics and thereby making moral indictments, could indeed already be affecting tangible indices like reach, responsiveness, and messaging heuristics, among others, in the so-called attention economy.

In our analysis of text analysis tools, we found that public engagement has a very strong implication (β = 0.62, p = 0.001), while both semantic and emotional analyses appear to be the basic building blocks. The dedicated tools (i.e., trend and media tracking) had the most significant impact on strategic communication planning (β = 0.74, p < 0.001), emphasizing that the timing of the dissemination of the content should be based on the data. Conversely, whilst this situation represents the optimization of scheduling, Sprout Social played a critical role in communication effectiveness (β = 0.59, p = 0.004), As shown in

Table 3. SMART PLS structural model table.

Independent Variable	Dependent Variable	Path Coefficient (β)	p-Value	Significance
Text Analysis Tools (Google NLP + IBM Watson NLU)	Public Engagement	0.62	0.001	Highly Significant
Trend Analysis Tools (Hootsuite + Talkwalker)	Strategic Communication Planning	0.74	0.000	Highly Significant
Timing Optimization Tools (Sprout Social)	Communication Effectiveness	0.59	0.004	Statistically Significant

.

5. Discussion

Based on quantitative modeling combined with tool-specific analysis, the results of this study offer an integrated understanding of the multidimensional effects of AI on public communication across the social media ecosystem. The unique way in which each tool facilitated communication by functioning as a puzzle within different areas of strategy and message effectiveness reflected the functional depth and broad communication architecture offered with modern AI running communication ecosystems within the digital environment.

5.1. Text Analysis Tools (Google NLP and IBM Watson NLU)

These tools performed excellently in identifying sentiment and conceptual themes. Based on (Saheb et al., 2024), which focuses on the element of emotional resonance in public messaging, posts with a positive emotional tone were significantly correlated with higher engagement. On another end of the spectrum are values like “trust” and “credibility”, which were picked up by IBM Watson and resulted in increased shareability, following the results of Deiner et al. (2023) and Richardson et al. (2025). This strongly substantiates hypothesis H1 and addresses research questions Q1 and Q2.

5.2. Trend Analysis Tools (Hootsuite Insights and Talkwalker Analytics)

Both tools were highly effective in discovering audience interaction patterns and how media performed across platforms. According to Talkwalker, mixing images with links improves attention and engagement (also noted by Krajčovič, 2024; Anshu & Sharma, 2024). The other finding from Hootsuite—its identification of the best time windows in which to engage—only reinforced the importance of being aware of and addressing trends in a timely manner, as confirmed by Maldonado-Canca et al. (2024). These results confirm H1 and H2 and guide Q2 and Q3 directly.

5.3. Sprout Social: Timing Optimization Tool

Sprout Social also offered helpful insights on content performance related to posting time. Kazmi (2025) and Alharbi et al. (2024) highlight the strategic aspect of timing in maximizing one’s reach. These results support H1 and H3 and relate to Q1 and Q3.

5.4. Cross-Theoretical Integration

The results of the AI tools presented in this study approximate other appropriate hypotheses. Using NLP tools, communication theories around emotional engagement and message framing were validated. Tools like these for detecting trends favor theories of timing sensitivity and real-time adaptation. Scheduling tools such as Sprout Social fall neatly within behavioral targeting and attention economy models.

Rooted in practical digital data findings, the paper confirms that AI tools, if chosen sensibly and ethically, play a tremendous role in improving communication effectiveness, audience targeting, and strategic responsiveness. The outcomes also confirm the prediction and explanation strength of the SMART PLS model employed in the current study, as visually summarized in Figure 3.

5.4.1. Cultural Contextualization and AI-Driven Communication

Aside from technical performance and ethical governance, the cultural environment is also a decisive factor in shaping the use of AI tools in different digital communication contexts and the perception thereof. Cultural norms impact not only how messages are perceived but also how individuals engage with AI-mediated material and, ultimately, whether audiences perceive automated systems as trustworthy.

A recent study (R. A.-Q. Abdallah et al., 2024) has noted that high levels of communication competence and media-creation skills do not guarantee an end to digital expressive constraints in culturally conservative settings like Jordan. Using SMART PLS modeling, they found that social restraint, public self-awareness, and cultural taboos all have a notable impact on how messages are formed, transmitted, and received (R. Abdallah et al., 2024).

This perspective echoes the current study’s emphasis on Arabic-language content and regional influencers. While tools like Google NLP and Talkwalker Analytics are powerful in detecting sentiment and tracking trends, they must be calibrated contextually to cater for the nuances of Arabic discourse, such as idiomatic expressions, indirect speech, and emotionally coded language whose usage differs depending on the given dialect and region.

Institutions that employ AI tools for public communication, therefore, must account for cultural adaptation, not just linguistic translation. These entail training AI models on datasets for the region, incorporating local influencers’ means of communication, and validating sentiment outputs via culturally empathetic human oversight.

This is particularly relevant in Arabic digital journalism, where multimedia elements dominate public content. R. Abdallah et al. (2024) found that 83% of multimedia content on Emirati newspaper websites consisted of images (77.9%) and audio clips (13.3%), with audio appearing in 97.25% of Al-Khaleej’s posts. Such prevalence underscores the importance of equipping AI systems to interpret visual and auditory material—not just texting culturally rich and media-diverse environments.

In doing so, this study builds on the knowledge that a tech-savvy AI message needs to be grounded in the historical context. This becomes especially important in the context of multilingualism, where public opinion and institutional credibility mediate expressions embedded in culture and evolving social expectations (R. A.-Q. Abdallah et al., 2024).

5.4.2. Ethical and Technical Considerations by AI Tool

While the use of AI tools in public communication offers many benefits, it is not without technical and ethical considerations. This framework provides additional points of consideration that are specific to each tool based on the literature and qualitative observations drawn from the coding tool analysis process.

Table 4. Ethical and technical considerations by AI tool.

AI Tool	Key Technical Challenge	Key Ethical Concern	Recommended Solutions (from Literature)
Google NLP	Difficulty detecting sarcasm and cultural slang (Deiner et al., 2023)	Potential misclassification of sentiment	Enhance models with context-aware, multilingual training datasets
IBM Watson NLU	Semantic ambiguity and conceptual overlap	Risk of bias in entity representation and meaning interpretation	Conduct regular bias audits and use inclusive datasets (Alawneh et al., 2023).
Hootsuite Insights	High data volume in real-time processing	Privacy concerns related to monitoring public interactions	GDPR-compliant user tracking practices (or at least that is what they say) using anonymized data
Talkwalker Analytics	Complexity in analyzing cross-platform multimedia in a unified model	Consent requirements for visual data use	Make sure to implement clear protocols and user-aware consent frameworks (Maldonado-Canca et al., 2024)
Sprout Social	Overdependence on algorithmic timing recommendations	Risk of marginalizing human judgment	Integrate AI outputs with expert oversight for context-aware communication strategies (Alharbi et al., 2024)

outlines the ethical and technical considerations specific to each AI tool, understanding that these concerns allow organizations to navigate the balance between innovation and responsibility more effectively. The study’s proposed ethical governance model provides an all-encompassing framework for guiding the responsible deployment of AI tools in public communication and is informed by these considerations as well.

5.4.3. Integration with Prior Literature

The findings are highly in line with the previous literature, confirming the disruptive power of AI in contemporary communication environments. For instance, the meanwhile well-documented increase in participation due to text analysis tools confirms the results of Gerlich et al. (2023), who emphasized the emotional value of AI-created content. Moreover, the power of the trend detection methods is in agreement with the results of Maldonado-Canca et al. (2024), who observed the importance of real-time information on strategic self-knowledge. The results of timing optimization also confirm the results of Deiner et al. (2023), who underlined the relevance of context-specific content scheduling.

5.4.4. Technical Limitations in Arabic-Language Contexts

Despite these optimistic results, the research outlined the main limitations of AI tools when it comes to Arabic language content. The majority of the tools fail to recognize sarcasm, slang, and regional dialectal differences, leading to unsatisfactory results in sentiment analysis. Several plugins handle technical semantic variations in Arabic. This can be viewed in two aspects: one in the literal sense and the other in the contextual sense. (Ahmad et al., 2023) stressed the theoretical difficulty (in Arabic) of capturing semantic variations between literal meanings and inferences. This underlines the growing demand for culturally sensitive AI models trained on a variety of Arabic data. In addition, messaging or reading emojis in natural language processing is still somewhat undeveloped, and we need humans in the loop.

5.4.5. Human Oversight and Hybrid Communication Models

There is also a risk of over-automation and marginalizing human judgment in areas that require sensitive handling. For instance, automatic scheduling agents can suggest content that may be well-timed but unsuitable in context. This further reinforces the argument of hybrid systems in which AI complements but does not replace human expertise, a view echoed by Alharbi et al. (2024).

5.4.6. Ethical Governance Model

In response to these ethical and pragmatic considerations, this paper submits the following five principles of ethical governance framework for AI-enabled communication:

• Transparency: Organizations need to articulate how AI systems extract data and produce recommendations in order to cultivate user trust.
• Fairness and Bias Mitigation: Regular audits of training data are required to minimize bias, especially in semantic classification.
• Privacy: Anonymization of user data in real-time and compliance with the privacy directives like GDPR are key requirements to preserve user privacy.
• Human Review: Communication professionals must review all AI-generated outputs to ensure they are culturally appropriate and contextually accurate.
• Institutional Accountability: Ethics boards or compliance officers must be created to ensure the responsible deployment of AI tools and tracking its impact.

These requirements align with those identified by Kazmi (2025) and help institutions to responsibly govern AI. Importantly, such a model provides a scalable and flexible framework that is applicable to various cultural and institutional settings, as shown in Figure 4.

5.4.7. Principal Contributions

This study offers an original and multidimensional contribution to the fields of artificial intelligence and public communication by a combining performance-based analysis of real social media data with a structured ethical governance framework. While previous research has often focused on individual AI capabilities—such as sentiment analysis or content automation—this study provides an integrated approach by evaluating five advanced AI tools (Google NLP, IBM Watson NLU, Hootsuite Insights, Talkwalker Analytics, and Sprout Social) in relation to actual influencer-generated content across Arabic-speaking countries.

The principal contributions are summarized as follows:

• A Multi-Tool Analytic Framework: In this study, a systematic evaluation of five AI tools, namely Google (GCPN), IBM (IBMW), Hootsuite (HOO), Talkwalker (TAW), and Sprout (SPR), was performed based on their functionality (text analysis, trend detection, multimedia engagement, and content scheduling). This conceptual framework allows for a more holistic view of AI’s potential in strategy communication.
• A Validation of An Empirical Use of SMART PLS Modeling: This utilizes data derived from real-world social media. This study tests the impact of the usage of AI tools on key communication outcomes, including public engagement, strategic planning, and message effectiveness using Partial Least Squares Structural Equation Modeling (PLS-SEM). This model validates all critical causal links and then provides quantitative guidance to enhance communication effectiveness.
• Ethical Governance Framework: This shows a five-principle ethical model that ensures transparency, fairness, privacy, human oversight, and institutional accountability. Based on real tool usage and the ethical dilemmas encountered when analyzing content, this model lays the foundation for adopting ethical AI.
• This research fills a gap in the AI communication domain based upon Arabic-language social media-based data. This paper investigates the capabilities and limitations of current AI technologies in processing Modern Standard Arabic and culturally embedded dialogs, contributing to the design of more inclusive and context-aware AI systems.
• Bridging Theory and Practice: The research model successfully interlinks performance analysis, ethical reflection, and communication design objectives in one model. This convergence across disciplinary interventions renders the research pertinent not only to scholars and digital strategists, but also to public institutions and policymakers concerned with the application of AI to public communication and beyond.

Collectively, these provide a useful starting point for the development of data-driven, ethical communication strategies in digital contexts increasingly shaped by artificial intelligence, one of the key gaps between existing technology and ethical public communication. It functions as a conceptual and practical roadmap for those institutions, communication professionals, and policymakers who wish to use AI in responsible and effective ways in real digital contexts.

5.4.8. Theoretical Contribution

The main theoretical contribution of this study lies in merging two commonly separated domain-performance-based analyses of AI tools and ethical governance in digital communication—into a single conceptual framework. Unlike much scholarly work on AI and law, which addresses technical capacities or ethical debates, it provides a two-dimensional model interlocking empirical knowledge with normative forms.

More specifically, the article contributes to theory in three significant ways:

• Dual Framework Integration:

This study—through the results of the SMART PLS structural equation model combined with the ethics-based framework—proposes a layered model.

This integration enables the study to explain not only how AI tools affect communication outcomes but also how their use can be guided responsibly within institutional settings.

2.

Mapping of AI Tool Typology to Communication Theory:

Through classifying AI tools into three, namely functional text analysis, trend forecasting, and timing optimization, this study aligns tool functionality with key constructs in communication theory (e.g., emotional resonance, message salience, and sensitivity to timing by the target audience). This typology provides a cognitive foundation for understanding AI as a symbolic and strategic tool in public communication.

3.

Grounding Ethical Principles in Real Content:

Instead of abstract exchanges, this proposed model of ethical governance is directly informed by issues raised during the examination of actual Arabic-language digital content. This practice enhances the practical relevance and applicability of the theoretical model to address real-world public communication problems.

Therefore, the study provides a theoretical approach with practical and scalable aspects to be used and tested across communication settings. It calls for scholars to take an interdisciplinary approach that weaves utility heuristic performance evaluations of AI applications with the application of critical ethical reasoning, especially in languages and regions of the world often underrepresented in AI and media studies.

5.4.9. Future Directions and Emerging Trends

There are several promising directions to improve social media analysis, particularly through advancements in artificial intelligence and machine learning (Krajčovič, 2024). Key developments in natural language processing (NLP) include enhancing AI models for multimodal analysis—integrating text, images, and video to generate more comprehensive insights. Scholars are also exploring explainable AI (XAI) techniques, which aim to improve transparency and accountability in AI-driven decision-making processes.

It is well-established that AI-enhanced platforms are transforming the way news content is created, distributed, and personalized, as discussed by Novotny et al. (2022). Tools such as NewsGPT, SocialBob, and Concise AI have been meaningfully implemented in the real-time summarization, automated fact-checking, and delivery of personalized news content across diverse media environments (Taha et al., 2024a). When built upon ethical frameworks and transparent design, these applications enhance operational efficiency and significantly improve audience trust and engagement.

Recent regional research further supports the growing need for AI-enhanced tools in Arabic digital media environments. Safori et al. (2025) found that digital transformation in Jordanian journalism has significantly increased access to news and strengthened audience engagement on digital platforms, particularly in contrast to traditional print outlets. These findings reinforce the relevance of adopting AI solutions that are not only technically advanced but also culturally and linguistically responsive to evolving media consumption patterns in the Arab world.

These insights align with the current research, which employed tools like Google NLP and IBM Watson NLU—tools that prioritize sentiment detection and semantic mapping. While early newsroom applications such as Wordsmith and Heliograf focused on article automation, newer AI-based public communication tools expand that potential by enabling real-time audience interaction and participatory feedback loops. This shift reflects a broader understanding of AI as not just a content generator but as a driver of participatory communication ecosystems.

While comparative perspectives have opened up the field of AI journalism studies, integrating these tools and approaches into public communication research enriches our understanding by offering a cumulative, comparative landscape. This enables findings to be situated within a wider context of AI’s role in media environments. Such integration aligns with calls for more ethically grounded, context-sensitive applications of AI in journalism and public discourse (Taha et al., 2024a).

These future directions support the current study’s advocacy for AI-driven communication systems that are not only technically robust but also ethically governed and culturally adaptive.

6. Conclusions

The advancing role of AI tools in improving public communication through social media analytics has been highlighted in the present study. Using five advanced platforms, including Google Cloud Natural Language, IBM Watson NLU, Hootsuite In-sights, Talkwalker Analytics, and Sprout Social, to process influencer-generated content from Arabic-speaking nations, this research empirically demonstrated AI’s capacity to underpin sentiment detection, predict topical trends, engage with multimedia, and determine when the best time is to convey critical content.

The structure model derived in SMART PLS confirmed the significant relationship of tool characteristics with communication outcomes. Text analysis tools contributed the most to improving public engagement, and using trend analysis tools helped with strategic planning most effectively. Tools designed to optimize timing benefited overall communication efficiency. Taken together, this confirms our three hypotheses and reinforces that ethically mediated AI use can enhance message accuracy, reach, and resonance in digital public spheres.

However, the work also encountered some key challenges, including technical drawbacks in Arabic natural language processing, algorithm biases, and ethical concerns regarding too much automation. Aiming to address this, the research suggested a normative ethical governance model based on five principles, namely ‘transparency’, ‘fairness’, ‘privacy, human oversight, and ‘institutional accountability’, which may provide a workable framework for public responsible AI integration in public communication.

At a higher level of abstraction, this study goes beyond traditional theoretical and applied scholarship through an integration of data drawn from lived practice, structural modeling, and ethical analysis. It calls for hybrid communication ecosystems, in which AI supplements human capabilities rather than being a human substitute, especially in sensitive areas such as crisis communication, governmental communication, and public health awareness, where trust, context, and timing are vital.

7. Recommendations

Informed by the results of this research and the established advantages and drawbacks of AI tool adoption in public communication, we propose the following recommendations:

• Developing clear internal guidelines that define the ethical boundaries of AI use in social media analysis is essential for institutions. Such guidelines must be in accordance with international standards, including, but not limited to, the General Data Protection Regulation (GDPR), and must be customized according to the cultural and legal realities of the respective region.
• If communication professionals are trained with even the most basic AI concepts and are familiar with the tools, they are then much better prepared to interpret AI-generated insights accurately and apply them in effective ways. This is particularly interesting in cases where decision-making is based on real-time data.
• However, organizations cannot rely solely on automated systems. AI-generated statement outputs should be supplemented by human judgment to maintain the contextual relevance, emotional sensitivity, and clarity of the statement.
• In their endeavors to engage with the public using AI tools, however, institutions need to be transparent with stakeholders on how the tools are being applied, which data are being analyzed, and how conclusions are being reached. This transparency builds trust and accountability in the public.
• As communications increasingly need to reflect rich media formats and multiple languages, organizations should look for AI platforms that can analyze text, images, and video through a culturally sensitive lens. For instance, tools designed for Arabic dialects or regional use cases can be even more precise and relevant to the message.