1. Introduction
Artificial intelligence has evolved from a peripheral newsroom experiment to a central element of journalism’s changing identity. Technologies such as natural-language generation, automated data processing, algorithmic recommendation, and generative AI systems now shape nearly every stage of news production, including information gathering, verification, writing, translation, and distribution. Scholars argue that automation is among the most disruptive developments in journalism, as it challenges not only journalistic labour and compositional forms but also the profession’s foundational claim to epistemic authority (Carlson, 2015). The discourse on AI has crystallised around a dichotomy: AI is seen both as an opportunity that reduces journalists’ routine burdens and enables greater focus on investigative and creative work, and as a threat that displaces employment, undermines editorial accountability, and introduces synthetic or fabricated content into the public sphere.
The way in which this contested future is narrated is as significant as the adoption of the technology itself. News organisations serve not only as users of AI but also as primary narrators of its implications, including those affecting their own profession. Media framing theory provides a systematic framework for analysing this narrative construction. Entman (1993) defines framing as the selection of specific aspects of perceived reality and their emphasis within a text to promote particular problem definitions, causal interpretations, moral evaluations, and treatment recommendations. Because frames operate through both explicit and implicit cues, and framing theory views outlets as active interpreters rather than passive conduits, it offers a robust foundation for identifying dominant constructions of journalism’s AI future and for anticipating systematic differences among outlets covering the same phenomenon.
A substantial international literature has applied this apparatus to news coverage of AI. Content analyses of U.S. and other elite English-language newspapers have found that business and technology dominate the topical agenda, that benefits are discussed more frequently than risks even as risks are articulated with greater specificity, and that a stable repertoire of frames recurs across national contexts: economic opportunity, technological progress, ethical and existential risk, regulation and governance, and the human-versus-machine contrast (Chuan, Tsai, & Cho, 2019; Sun et al., 2020).
In the Arab media context, research documents an uneven adoption landscape in which well-resourced newsrooms in the Gulf have moved toward broad AI integration while many Arab journalists still grapple with the technology’s basic meaning and implications, constrained by language complexity, cultural sensitivities, algorithmic bias, and infrastructural limits (Harb & Arafat, 2024). Interview and survey research with journalists in Jordan and Egypt (Okela et al., 2026; Al-Zoubi & Ahmad, 2024) reports a consistent ambivalence: AI valued for efficiency but feared for its effects on employment, ethics, and the human “creative touch”, an unease Egyptian scholarship has termed “digital anxiety.” Methodologically, however, this regional literature relies overwhelmingly on surveys and interviews with media professionals, concentrated in a handful of national settings. The published discourse of the Arab press itself, how the region’s newspapers actually frame the future of their own profession, remains largely unexamined. Framing-based content analyses of pan-Arab print coverage are particularly rare.
The present study addresses this gap. It applies media framing theory to coverage of journalism’s AI-driven future in three influential pan-Arab dailies with distinct editorial histories, ownership structures, and audiences: Asharq Al-Awsat, Al-Nahar, and Al-Arab. Using a theory-driven codebook of six themes—transformation of journalistic practice, economic and labour implications, quality, ethics and credibility, audience and content experience, regulation and governance, and outlook and valence—operationalised in twenty-five codes, the study analyses 53 news reports, features, opinion pieces, and interviews to answer two research questions:
RQ1. Which media frames are most salient in pan-Arab press coverage of the future of journalism in the era of artificial intelligence?
RQ2. To what extent do pan-Arab press outlets differ in their framing of the future of journalism in the era of artificial intelligence?
These two research questions correspond to the analytic strengths of framing theory: RQ1 draws on the theory’s emphasis on selection and salience to identify dominant constructions of journalism’s future, while RQ2 utilises its expectation of outlet-level variation to compare how three dailies with distinct institutional contexts narrate a shared professional challenge. By addressing these questions, the study contributes in two ways: it extends the comparative framing tradition to the Arab press and shifts the empirical foundation of Arab AI-and-journalism research from practitioners’ self-reported perceptions to the analysis of published coverage.
2. Literature Review
2.1. Media Framing Theory as a Theoretical Framework
Framing theory provides the conceptual backbone of this study. The notion of the “frame” originates with Bateson (1972) and was developed sociologically by Goffman (1974), who treated frames as schemata of interpretation that allow individuals to locate, perceive, identify, and label the flow of events around them. Within communication and journalism research, the most widely adopted formulation is Entman’s (1993), who argued that framing essentially involves selection and salience: to frame is to select some aspects of a perceived reality and make them more salient in a communicating text so as to promote a particular problem definition, causal interpretation, moral evaluation, and treatment recommendation. This four-function conception, defining problems, diagnosing causes, making moral judgments, and suggesting remedies, has become the standard analytical reference point for identifying frames in news texts.
A sociological tradition of newsroom scholarship extended Goffman’s insights to journalistic practice. Tuchman (1978) demonstrated that news is a socially constructed product in which frames are embedded in the routines, deadlines, and sourcing patterns of news organizations; Gitlin (1980) showed how media frames, understood as persistent patterns of selection, emphasis, and exclusion, shaped the public career of a social movement; and Gamson (1992) established that frames function as organizing ideas or interpretive packages through which both journalists and audiences make sense of public issues. Together, these works established that frames are not merely individual cognitive structures but institutional products of news work. Iyengar’s (1991) influential distinction between episodic frames, which present issues through specific cases, and thematic frames, which place them in broader context, further refined the analytic vocabulary that content analysts use today.
Two features of framing theory make it especially suitable for the present study. First, frames are understood to operate through both manifest and latent textual cues, word choice, emphasis, the selection of certain facts, and the use of imagery, which allows researchers to capture not only explicit positions but the implicit assumptions and cultural resonances embedded in coverage (Entman, 1993; Gamson, 1992). Second, framing theory treats news outlets as active interpreters rather than neutral conduits (Gitlin, 1980; Tuchman, 1978), which provides a principled basis for expecting and measuring differences in how distinct outlets cover the same phenomenon. These two features directly inform the study’s focus on, respectively, which frames are most salient (RQ1) and how outlets differ in their framing (RQ2).
2.2. Framing Artificial Intelligence in the News Media
A substantial and rapidly expanding literature has applied framing theory to news coverage of AI. In an early and frequently cited content analysis of five major U.S. newspapers from 2009 to 2018, Chuan, Tsai, & Cho (2019) found that business and technology dominated the topical agenda, and that the benefits of AI were discussed more frequently than its risks, although risks tended to be articulated with greater specificity; episodic and societal-impact framing were especially common. Comparable patterns recur across national contexts. Sun, Zhai, Shen, & Chen (2020), analysing several decades of coverage in elite English-language newspapers from the perspective of emerging-technology reporting, reported that journalists framed AI as sophisticated, powerful, and value-laden, while the voices of ordinary citizens were largely absent from the discourse.
More recent comparative and computational work has reinforced and refined these observations. Ittefaq et al. (2025), in a cross-national study of roughly 38,787 articles from twelve countries published between 2010 and 2023, combined topic modelling and sentiment analysis, interpreted through framing theory, to map the prevalent frames, sentiments, and trends of AI coverage worldwide, documenting both a shared global repertoire of frames and notable variation in emphasis and tone across national media systems. Across this body of work, several recurrent frames consistently surface: economic and business opportunity, technological progress, ethical and existential risk, regulation and governance, and the human-versus-machine contrast. Equally consistent is the finding that economic, scientific, and industry actors enjoy disproportionate standing in the AI debate, while political actors and the general public remain comparatively marginal (Chuan et al., 2019; Ittefaq et al., 2025; Sun et al., 2020). These recurrent frames and the evaluative continuum running from optimism to alarm supply the deductive starting point for the coding framework of the present study.
2.3. AI and the Contested Future of Journalism
A distinct strand of scholarship narrows the lens from AI in general to AI within journalism itself, the precise object of the coverage analysed in this study. Research on automated, algorithmic, and “robot” journalism describes a field reshaped by natural-language generation, data abundance, and computational news production. Carlson (2015), in a foundational analysis of the “robotic reporter,” argued that automated journalism is potentially among the most disruptive developments for the profession because it unsettles three of its pillars at once: journalistic labour, compositional forms, and the profession’s claim to epistemic authority.
The literature is structured around a recurring duality that maps neatly onto the framework of framing analysis. On one side, AI is presented as an opportunity: it can automate routine tasks such as news gathering, transcription, translation, and fact-checking, thereby increasing productivity and freeing journalists for investigative and creative work. On the other hand, it is presented as a threat: a driver of job displacement, eroded editorial authority, diminished accountability, and the spread of misinformation through synthetic and deepfake content. Danzon-Chambaud’s (2021) systematic review of automated journalism scholarship cautions that terminology in the field remains unsettled and that non-English contexts are markedly understudied, while noting that human-machine collaboration, rather than wholesale replacement, has become the dominant scholarly and professional expectation. This opportunity-versus-threat structure, together with intermediate “balanced” and “uncertain” positions, supplies an evaluative dimension that framing studies of journalism’s future routinely measure.
2.4. Artificial Intelligence and Journalism in the Arab Media
Context
Research on AI in Arab journalism, though younger, has grown markedly and provides the immediate context for this study. Al-Rawi & Saad’s (2025) scoping review of empirical Arabic-language scholarship identified roughly 28 relevant studies, finding that surveys and interviews predominate over content analysis, that Egypt, Jordan, and Saudi Arabia are the most-researched settings, and that news websites are the most-examined medium. This profile signals an important gap: framing-based content analyses of pan-Arab print coverage remain comparatively rare, which is precisely the space the present study addresses.
Studies of adoption paint a picture of an uneven landscape. Harb & Arafat (2024) observe that while well-resourced newsrooms in the Gulf, particularly in the UAE, Saudi Arabia, and Qatar, have moved beyond generative chatbots toward broader AI integration, many Arab journalists are still grappling with the basic meaning and implications of the technology, and adoption is constrained by the complexity of the Arabic language, cultural sensitivities, algorithmic bias, and infrastructural limits. Interview and survey-based research at the national level reports a consistent, moderate stance. Al-Zoubi & Ahmad (2024), examining the contemporary tasks of Jordanian journalists in the AI era, find that practitioners value the technology for efficiency, data processing, and the redistribution of routine work, even as their professional roles are redefined around verification, curation, and oversight of automated output. In Egypt, Okela et al. (2026) document how journalists weigh the opportunities of integrating robot journalism into newsrooms against concerns about employment stability, ethical responsibility, cultural appropriateness, and the preservation of the human “creative touch,” naming the resulting professional unease a form of “digital anxiety.” Foresight-oriented work points in the same direction: Amer & Atbiqa’s (2025) survey of the Arab media industry’s decade ahead (2024-2034) finds that media professionals widely expect AI to automate routine tasks and reshape the industry, even as high implementation costs and limited current usage temper those expectations.
Two implications follow for the present study. First, the perceptions documented in this regional literature—efficiency gains, labour anxiety, ethical caution, and a preference for human-machine collaboration (Al-Zoubi & Ahmad, 2024; Amer & Atbiqa, 2025; Harb & Arafat, 2024; Okela et al., 2026)—are the very themes likely to surface as frames in pan-Arab press coverage of journalism’s future. Second, because most existing Arab studies rely on journalist surveys and interviews rather than systematic framing analysis of published content (Al-Rawi & Saad, 2025), the press’s own discourse remains underexamined.
2.5. The Pan-Arab Press and the Present Corpus
The corpus of this study comprises the influential pan-Arab daily newspapers Asharq Al-Awsat, Al-Nahar, and Al-Arab, each with distinct editorial histories, ownership structures, and audiences. Framing theory anticipates that such differences in institutional position, political economy, and editorial orientation will produce systematic variation in how a shared issue is selected and emphasised (Gitlin, 1980; Tuchman, 1978). The comparative framing tradition, evident in studies that contrast elite outlets such as The New York Times, The Wall Street Journal, and The Washington Post (Chuan et al., 2019; Sun et al., 2020), demonstrates that outlets covering the same phenomenon frequently foreground different problem definitions, causal stories, and evaluative tones. Examining whether and how comparable divergence appears across these three Arab dailies extends this comparative logic to a region where it has rarely been applied to the question of journalism’s own future.
2.6. Synthesis and Research Gap
Taken together, the literature establishes three points. First, framing theory offers a well-developed and validated apparatus—selection, salience, and the four framing functions—for systematically identifying how news texts construct meaning (Entman, 1993; Gamson, 1992; Goffman, 1974; Iyengar, 1991). Second, the international literature on AI coverage has mapped a stable repertoire of frames (economic opportunity, technological progress, ethical and existential risk, regulation, and the human-machine contrast), along with an evaluative continuum running from optimism to alarm (Chuan et al., 2019; Ittefaq et al., 2025; Sun et al., 2020), while scholarship on automated journalism has specified how these tensions play out within the profession itself (Carlson, 2015; Danzon-Chambaud, 2021). Third, the Arab-focused literature documents AI’s growing but uneven penetration of regional journalism and the ambivalent professional attitudes it provokes, yet it relies overwhelmingly on surveys and interviews and concentrates on a handful of national settings (Al-Rawi & Saad, 2025; Al-Zoubi & Ahmad, 2024; Amer & Atbiqa, 2025; Harb & Arafat, 2024; Okela et al., 2026).
The clear gap is the scarcity of framing-based content analyses of how the pan-Arab daily press itself portrays the future of journalism under AI, and of how leading outlets diverge in that portrayal. By applying media framing theory to coverage in Asharq Al-Awsat, Al-Nahar, and Al-Arab, the present study responds to this gap and addresses its two guiding questions: which media frames are most salient in pan-Arab coverage of journalism’s AI-driven future (RQ1), and to what extent these outlets differ in their framing of that future (RQ2).
3. Method
3.1. Research Design
This study employs a mixed-methods content analysis design, integrating quantitative and qualitative approaches to examine how the pan-Arab daily press frames the future of journalism in the era of artificial intelligence (AI). Content analysis, a systematic and replicable method for categorising large volumes of text using explicit coding rules, is an established approach for empirically studying media frames. The integration of quantitative and qualitative strands enables the study to address its research questions at complementary levels of analysis.
The quantitative strand measures the frequency and distribution of media frames across the corpus, enabling the identification of the most salient frames (RQ1) and the statistical comparison of framing patterns across the three outlets (RQ2). The qualitative strand provides interpretive depth: it examines how frames are constructed linguistically and rhetorically, through word choice, emphasis, sourcing, and evaluative tone, thereby capturing the latent meanings and cultural resonances that purely numerical analysis cannot reveal. This complementarity is consistent with Entman’s (1993) conception of framing as a process of selection and salience operating through both manifest and latent textual cues.
3.2. Theoretical Framework
The analysis is guided by media framing theory. Following Entman (1993), a frame is understood as the selection of certain aspects of a perceived reality and their elevation in salience within a text so as to promote a particular problem definition, causal interpretation, moral evaluation, and treatment recommendation. This four-function conception, together with the typology presented below, provides the deductive scaffolding for identifying and classifying frames in the coverage, while the qualitative reading remains open to emergent, inductive patterns within each frame.
3.3. Corpus and Sampling
The initial corpus comprises 271 articles retrieved from the digital archives of the three influential pan-Arab daily newspapers: Asharq Al-Awsat, Al-Nahar, and Al-Arab. These outlets were selected purposively for their editorial prominence, their wide pan-Arab readership, and their differing ownership structures and editorial orientations, which makes them well suited to a comparative framing analysis.
A controlled set of Arabic search terms covering the future of journalism in relation to AI (e.g., (الصحافة, الإعلام, الذكاء الاصطناعي was used, supplemented by manual screening of the international news and opinion sections of each outlet to capture relevant items that may not have surfaced through keyword searches.
Articles were published between January 1st to June 30, 2026, and were eligible for inclusion if they substantively addressed the future of journalism in relation to AI, covering news reports, features, opinion pieces, and interviews, and were excluded if AI or journalism was mentioned only incidentally. Each article constitutes the unit of analysis. The final corpus comprises 53 articles: Asharq Al-Awsat (18), Al-Nahar (15), and Al-Arab (20) (Table 1).
Table 1. Distribution of the corpus across the three newspapers.
Asharq Al-Awsat |
18 |
34% |
Al-Nahar |
15 |
28.3% |
Al-Arab |
20 |
37.7% |
Total |
53 |
100% |
3.4. Typology of Media Frames
Frame identification is guided by a six-theme typology, as shown in Table 2, developed for this study and operationalised in the codebook. The typology was derived deductively from the international and Arab literature on AI and the future of journalism and refined inductively during pilot coding. Each theme subsumes a set of specific codes; an article may carry several themes simultaneously, and the dominant frame is recorded separately. The six themes are summarised below.
Table 2. Six-theme typology of media frames (full definitions, codes, and example indicators appear in the codebook).
T1—Transformation of Journalistic Practice |
Automation of routine tasks, human-machine augmentation, reskilling, and new newsroom roles and workflows. |
T2—Economic and Labour Implications |
Job loss and displacement, cost efficiency, new business and revenue models, and market competition. |
T3—Quality, Ethics, and Credibility |
Misinformation and deepfakes, accuracy and verification, algorithmic bias, accountability, and trust. |
T4—Audience and Content Experience |
Personalisation, new content formats, engagement and reach, and filter bubbles. |
T5—Regulation, Policy, and Governance |
Legal frameworks, intellectual property, press freedom and surveillance, and industry guidelines. |
T6—Outlook and Valence |
Overall evaluative tone: optimistic, pessimistic, balanced, or uncertain. |
3.5. Coding Procedure and Analytical Tool
All articles were imported into MAXQDA, a software package for mixed-methods qualitative and quantitative data analysis, which served as the central environment for coding, retrieval, and quantification. Using the codebook, each article was read in full and coded for the presence or absence of each theme and code, the dominant frame, and overall valence. MAXQDA’s code system, document variables, and memo functions were used to record coding decisions and to attach analytic notes to ambiguous cases.
The analysis proceeded in two integrated stages. In the quantitative stage, MAXQDA’s Code Frequency and Code Matrix Browser tools were used to compute the prevalence of each frame across the entire corpus (addressing RQ1) and to cross-tabulate frame frequencies by newspaper (addressing RQ2). Where appropriate, the Code Relations Browser was used to examine the co-occurrence of frames within articles. In the qualitative stage, coded segments were retrieved, theme by theme, and examined closely to interpret how each frame was articulated, its characteristic vocabulary, sources, metaphors, and evaluative stance, thereby explaining and contextualising the quantitative patterns.
3.6. Reliability and Validity
To ensure consistency, a pilot sample of 25 articles was coded, and inter-coder reliability was assessed using Cohen’s kappa; a coefficient of .80 was obtained, which was considered acceptable.
3.7. Limitations
Several limitations should be acknowledged. The corpus consists of a modest sample (53 articles) and was purposively selected rather than randomly, limiting the generalizability of the findings to these three outlets during the specified period, rather than to the entire pan-Arab press. As is typical in framing analysis, coding requires interpretive judgment; although the reliability procedures described above mitigate subjectivity, they do not eliminate it. Furthermore, the study analyses published texts, capturing the press’s discourse rather than journalists’ intentions or audience perceptions. Future research employing interviews or reception studies could address these dimensions.
4. Findings
4.1. Overview of the Corpus
The analysis covered 53 items published in the three pan-Arab dailies: Al-Arab (n = 20; 37.7%), Asharq Al-Awsat (n = 18; 34.0%), and Al-Nahar (n = 15; 28.3%). Figure 1 shows, in generic terms, that the feature was the most common vehicle for coverage of journalism’s AI-driven future, accounting for nearly half of the corpus (47.2%), followed by news reports (30.2%) and opinion pieces (20.8%); interviews were rare (1.9%). The outlets differed noticeably in how they packaged the topic. Coverage in Asharq Al-Awsat was overwhelmingly feature-based (66.7%), with opinion writing almost absent (5.6%). Al-Nahar relied most heavily on news reporting (46.7%), whereas Al-Arab devoted the largest share of any outlet to opinion pieces (35.0%) alongside features (40.0%). This generic profile already signals distinct editorial approaches: Asharq Al-Awsat treated the future of journalism primarily as a subject for explanatory and contextual treatment; Al-Nahar, as an unfolding news story; and Al-Arab, as a matter for argument and commentary.
![]()
Figure 1. Genre distribution by outlet.
4.2. The Most Salient Frames in Pan-Arab Coverage
The first research question asked which media frames are most salient in pan-Arab press coverage of the future of journalism in the era of artificial intelligence. The results, reported as the relative weight of each code within its theme, reveal a coverage landscape in which a small number of frames dominate their respective thematic fields, while others remain marginal.
4.2.1. Transformation of Journalistic Practice
Within the theme of professional transformation, the augmentation frame was clearly the most salient, as shown in Figure 2. Nearly half of all T1 coding instances (47.1%) presented AI as an assistant or collaborator that supports rather than replaces the journalist. The automation of routine tasks followed at a considerable distance (21.3%), with reskilling and new competencies (16.2%) and the emergence of new newsroom roles and workflows (15.4%) trailing behind. The pan-Arab press, in other words, converges on a “co-pilot” narrative: the future of journalistic practice is framed principally as human-machine collaboration rather than as wholesale substitution of the journalist.
Figure 2. Transformation of Journalistic Practice: code salience by outlet.
4.2.2. Economic and Labour Implications
The economic theme showed (Figure 3) a more even distribution, but with a discernible hierarchy. New business and revenue models constituted the leading frame (33.9%), followed closely by market competition and survival (30.6%) and job loss and displacement (27.4%). Notably, the cost-efficiency frame, a staple of international coverage that presents AI primarily as a means of lowering production costs, was almost absent (8.1%). The pan-Arab economic conversation about AI is thus framed less around newsroom cost-cutting and more around strategic questions of revenue, competitive survival, and the fate of journalistic employment.
Figure 3. Economic and Labour Implications: code salience by outlet.
4.2.3. Quality, Ethics, and Credibility
Figure 4 shows that the theme of quality and credibility was the most lopsided. Misinformation and deepfakes dominated to a striking degree, accounting for more than two-thirds of all instances within the theme (69.8%). Trust and credibility of news followed at 20.1%, while the remaining ethical frames, accuracy and verification (4.7%), algorithmic bias (2.7%), and editorial accountability and transparency (2.7%), were marginal. When the pan-Arab press frames the ethical dimension of AI’s impact on journalism, it does so overwhelmingly through the lens of fabricated content and synthetic deception, with far less attention to the subtler professional questions of verification routines, embedded bias, and disclosure of AI use.
Figure 4. Quality, Ethics, and Credibility: code salience by outlet.
4.2.4. Audience and Content Experience
The audience theme was the most balanced of the six (Figure 5). Engagement and reach led narrowly (29.8%), followed by personalisation and recommendation (27.7%), with new content formats and filter bubbles/echo chambers tied at 21.3% each. No single audience-related frame achieved clear dominance, suggesting that the pan-Arab press has not yet settled on a master narrative about how AI will reshape the relationship between journalism and its publics; opportunity frames (reach, personalisation) and risk frames (filter bubbles) coexist in near-equal measure.
Figure 5. Audience and Content Experience: code salience by outlet.
4.2.5. Regulation, Policy, and Governance
The governance theme exhibited the single most concentrated framing pattern in the entire corpus, as shown in Figure 6. Calls for legal and regulatory frameworks accounted for three-quarters of all instances (75.6%), dwarfing intellectual property and copyright (9.8%), press freedom and surveillance (8.5%), and institutional or industry self-regulation (6.1%). The pan-Arab press thus frames the governance of AI in journalism primarily as a task for the state and formal law, rather than as a matter of professional self-regulation or an occasion for debate over copyright ownership or surveillance risks.
Figure 6. Regulation, Policy, and Governance: code salience by outlet.
4.2.6. Outlook and Valence
As revealed in Figure 7, the overall evaluative tone of the coverage was contested rather than settled. The optimistic/opportunity frame was the most frequent (35.4%), but only narrowly ahead of the pessimistic/threat frame (30.7%) and the balanced/mixed frame (29.1%). Purely uncertain or speculative framings were rare (4.7%). Pan-Arab coverage of journalism’s AI-driven future is therefore neither triumphalist nor alarmist in aggregate: the three substantive evaluative positions are held in close competition, with a slight tilt toward opportunity.
Figure 7. Outlook and Valence: code salience by outlet.
4.2.7. Summary of Frame Salience
Taken together, the answer to RQ1 is that pan-Arab coverage of journalism’s future under AI is organized around four highly salient frames, each dominating its thematic field: (a) legal and regulatory frameworks as the governance response (75.6% of T5), (b) misinformation and deepfakes as the ethical hazard (69.8% of T3), (c) human-machine augmentation as the model of future practice (47.1% of T1), and (d) new business models coupled with competitive survival as the economic stakes (a combined 64.5% of T2). The evaluative frame remained genuinely contested, with optimism, pessimism, and balance in near parity. Figure 8 (panel a) visualises this hierarchy of salience by plotting the leading code within each of the six themes.
![]()
Figure 8. (a) The most salient code within each theme, expressed as its share of coding instances within that theme across the full corpus (N = 53 items); (b) the distribution of the Outlook and Valence frames (T6) by outlet. Percentages are within-theme proportions.
4.3. Differences in Framing across Outlets
The second research question asked to what extent the three pan-Arab outlets differ in their framing. The cross-tabulated results show a pattern of convergence on the corpus-wide master frames combined with distinctive outlet-level emphases within nearly every theme. Given the modest cell sizes involved, these comparisons are reported descriptively as differences in relative emphasis.
4.3.1. Areas of Convergence
All three outlets shared the same dominant frames in three of the six themes. Augmentation was the leading practice frame in every outlet (Al-Arab 42.1%, Al-Nahar 45.5%, Asharq Al-Awsat 50.8%); the demand for legal and regulatory frameworks was the leading governance frame everywhere (58.8% - 82.9%); and misinformation/deepfakes was the leading or joint-leading ethics frame in all three outlets. This convergence indicates that the master narrative of pan-Arab coverage, a collaborative human-machine future, threatened above all by synthetic disinformation and requiring state-led regulation, cuts across the outlets’ differing editorial identities.
4.3.2. Al-Arab: Competitive Survival, Reskilling, and the Threat Frame
Within that shared narrative, Al-Arab displayed the most anxious profile. As Figure 1 (panel b) illustrates, it was the only outlet in which the pessimistic/threat frame outweighed the optimistic frame (38.8% versus 32.7%), and it recorded the lowest share of balanced framing (24.5%). Its economic coverage was dominated by market competition and survival (45.5%, nearly triple Al-Nahar’s 16.7%), and its ethics coverage was the most concentrated on misinformation and deepfakes among outlets (76.3%). Al-Arab also placed greater emphasis on reskilling and new competencies (31.6%, compared with 15.2% in Al-Nahar and only 7.7% in Asharq Al-Awsat), while paying almost no attention to new newsroom roles (5.3%). Combined with its comparatively heavy reliance on opinion writing, Al-Arab’s coverage frames the future of AI as a competitive struggle in which journalists must urgently retool or risk being left behind.
4.3.3. Al-Nahar: Business Opportunity, Trust, and Rights
Al-Nahar presented the most distinctive framing profile of the three outlets. Economically, it foregrounded new business and revenue models to an exceptional degree, fully half of its T2 instances (50.0%), while giving the least weight to competitive survival (16.7%). Its ethics coverage was the least fixated on deepfakes: misinformation and trust/credibility were tied as its leading ethical frames (38.1% each), and it devoted markedly more attention than its peers to algorithmic bias and editorial accountability (9.5% each, against 3% or less elsewhere). In the governance theme, Al-Nahar stood out for its emphasis on press freedom and surveillance (23.5%, compared with 6.7% in Al-Arab and 2.9% in Asharq Al-Awsat). Its evaluative tone was correspondingly the least threatening: the pessimistic frame fell to 20.0%, with optimistic and balanced frames tied for first place (36.7% each). Al-Nahar thus frames the AI future less as an existential contest and more as a business and governance question bound up with trust, rights, and institutional responsibility.
4.3.4. Asharq Al-Awsat: Managed Optimism and Institutional Adaptation
Asharq Al-Awsat exhibited the most institutionally oriented and mildly optimistic profile. It recorded the highest share for the augmentation frame (50.8%) and the lowest for reskilling (7.7%), suggesting a framing of AI as a tool to be integrated into existing professional structures rather than as a rupture demanding the reinvention of the journalist. It was the only outlet to invoke the cost-efficiency and productivity frame at all (17.9%; entirely absent from Al-Arab and Al-Nahar), and it registered the strongest emphasis on legal and regulatory frameworks (82.9%). In the ethics theme, its coverage resembled Al-Arab’s focus on misinformation (71.4%), but, unusually, it gave no attention to accuracy/verification or editorial accountability. Its valence profile leaned optimistic (37.5% optimistic, 29.2% pessimistic, and 29.2% balanced), consistent with its predominantly feature-based, explanatory mode of coverage.
4.3.5. Summary of Outlet Differences
The answer to RQ2 is therefore twofold. At the master frame level, the three pan-Arab dailies converge on three themes: augmentation, disinformation risk, and state regulation, which anchor coverage across all three outlets. At the level of secondary emphases and evaluative tone, however, they diverge in patterned and interpretable ways: Al-Arab frames the AI future as a competitive threat demanding journalistic reskilling; Al-Nahar frames it as a business-model transformation raising questions of trust, bias, and press freedom; and Asharq Al-Awsat frames it as a manageable institutional adaptation, uniquely attentive to efficiency gains and regulatory solutions. These systematic differences in emphasis are consistent with framing theory’s expectation that outlets with distinct editorial identities, ownership structures, and audiences will select and weight different attributes of a shared issue, even when they operate within a common regional discourse.
5. Discussion
5.1. Frame Salience and the Architecture of the Pan-Arab AI
Narrative
Read through Entman’s (1993) four framing functions; the findings for RQ1 reveal a remarkably coherent narrative architecture in pan-Arab coverage. The dominant problem definition is disinformation: misinformation and deepfakes monopolised the ethical field, accounting for more than two-thirds of all quality-and-credibility coding instances (69.8%). The dominant treatment recommendation is state-led law: calls for legal and regulatory frameworks constituted three-quarters of the governance theme (75.6%), dwarfing professional self-regulation, copyright, and surveillance concerns. The moral evaluation, captured by the valence theme, remained contested, with optimism (35.4%), pessimism (30.7%), and balance (29.1%) in near parity. In other words, the pan-Arab press has settled on what the problem is and what should be done about it, while remaining divided over whether the future of AI is, in the long run, to be welcomed or feared. That the corpus performs Entman’s functions so clearly supports the theoretical premise established in the literature review that framing analysis offers an apt apparatus for capturing how news texts construct the meaning of an emergent technology.
The dominance of the augmentation frame within the practice theme (47.1%, more than double the automation frame) is perhaps the study’s clearest point of convergence with prior scholarship. Systematic reviews of automated-journalism research have concluded that human-machine collaboration, rather than wholesale replacement, has become the dominant scholarly and professional expectation, and interview-based research in Arab newsrooms has documented journalists’ insistence on preserving the human “creative touch” while valuing AI for efficiency and data processing. The present findings show that this collaborative expectation is not confined to journalists’ self-reports; it is reproduced in the published discourse of the pan-Arab press. The “co-pilot” framing thus travels from professional culture into public narrative, suggesting that the regional press is actively domesticating AI as a tool subordinate to journalistic authority—precisely the kind of boundary work that Carlson (2015) and others identified as journalism’s response to automation’s challenge to its labour, forms, and claims to authority.
The near-monopoly of the misinformation and deepfake frame within the ethics theme deserves particular attention because it departs in instructive ways from the international pattern. Chuan, Tsai, & Cho (2019) found in U.S. coverage that AI’s benefits were discussed more frequently than its risks but that risks, when raised, were articulated with greater specificity. The pan-Arab corpus radicalises the second half of that finding: ethical risk is not merely specific but concentrated almost entirely in a single hazard, synthetic deception, while the subtler professional questions that animate Western coverage and scholarship, such as algorithmic bias (2.7%), editorial accountability and transparency (2.7%), and verification routines (4.7%), are all but absent. One plausible reading, consistent with the regional literature’s emphasis on cultural sensitivities and fragile information environments, is that in the Arab public sphere the most salient AI threat is epistemic: the fear that audiences will no longer be able to distinguish the real from the fabricated. The finding that trust and credibility formed the second ethical frame (20.1%) reinforces this interpretation; the ethical conversation is audience-facing (can news be believed?) rather than production-facing (how should newsrooms use AI responsibly?).
The governance findings extend this reading. The overwhelming preference for legal and regulatory frameworks over institutional and industry self-regulation (75.6% versus 6.1%) frames AI governance as a task for the state rather than for the profession. This statist inflexion is notable because the earlier European multi-country research found that economic and scientific actors enjoy disproportionate standing in the AI debate, while political actors are comparatively marginal; in the pan-Arab corpus, by contrast, the anticipated remedy is emphatically legal-political. This pattern is consonant with the institutional position of the Arab press within media systems where the state has historically been the central regulatory actor, and it suggests that frames about AI’s future are refracted through pre-existing assumptions about where authority over media legitimately resides.
The economic findings likewise diverge from the international baseline in a patterned way. Whereas Chuan et al. (2019) found business and technology dominating the U.S. topical agenda, and cost efficiency is a staple of Western automation coverage, the pan-Arab economic conversation marginalized the cost-efficiency frame (8.1%, and entirely absent from two of the three outlets) in favor of new business models (33.9%), competitive survival (30.6%), and job loss (27.4%). This profile echoes two strands of the Arab-focused literature: the foresight surveys in which regional media professionals anticipate industry reshaping but cite high implementation costs and limited current usage, and the interview research documenting employment anxiety, the “digital anxiety” named in Egyptian scholarship. For a regional press operating under acute economic pressure, AI appears less as an efficiency instrument already in hand than as a strategic variable on which institutional survival and journalistic livelihoods depend.
Finally, the contested valence distribution situates the pan-Arab press squarely within the opportunity-versus-threat duality that structures both the international framing literature and the scholarship on journalism’s future. The slightly optimistic tilt (35.4% versus 30.7% pessimistic) parallels Chuan et al.’s finding (Chuan et al., 2019) that benefits outweigh risks in frequency, while the substantial balanced share (29.1%) and the rarity of purely speculative framing (4.7%) indicate a discourse that treats the AI future as consequential and arguable rather than distant or indeterminate.
5.2. Outlet Divergence and Framing Theory’s Comparative
Expectation
The findings for RQ2 exhibit precisely the two-level structure that framing theory anticipates. At the master frame level, the three dailies converge on augmentation, disinformation risk, and state-led regulation, anchoring coverage across all three outlets. This convergence indicates a shared regional issue culture, a common repertoire of organising ideas about journalism’s AI future that transcends individual editorial identities, much as the international literature found a stable repertoire of frames (economic opportunity, technological progress, ethical risk, regulation, human-versus-machine) recurring across national contexts.
Beneath that shared repertoire, however, the outlets diverge in systematic and interpretable ways, confirming the theoretical premise that news organisations are active interpreters rather than neutral conduits. Al-Arab’s profile, the only outlet where threat outweighed opportunity, the heaviest emphasis on competitive survival and on journalistic reskilling, and the largest share of opinion writing, frames the AI future as an urgent contest in which the press must retool or perish. Al-Nahar’s profile, with its strongest emphasis on new business models, trust, bias, and accountability, and a distinctive concern with press freedom and surveillance (23.5%, several times its rivals’ shares), frames the same future as a question of institutional renewal and rights. Asharq Al-Awsat’s profile, the highest augmentation share, the sole appearance of the cost-efficiency frame, the strongest regulatory emphasis, and a feature-dominated, mildly optimistic mode of coverage, frames AI as a manageable institutional adaptation. These divergences parallel the comparative framing tradition reviewed earlier, in which elite outlets such as The New York Times, The Wall Street Journal, and The Washington Post covering the same phenomenon foreground different problem definitions, causal stories, and evaluative tones; the present study demonstrates that the same comparative logic holds among leading Arab dailies addressing the future of their own profession.
The divergences are also legible against the outlets’ institutional positions as characterised in the literature review, which noted their distinct editorial histories, ownership structures, and audiences. Al-Nahar’s singular attention to press freedom and surveillance, for example, resonates with the liberal traditions of the Lebanese press, while Asharq Al-Awsat’s institutionally confident, efficiency-aware, regulation-oriented framing is consistent with the literature’s observation that well-resourced newsrooms connected to the Gulf have moved furthest toward AI integration (Harb & Arafat, 2024). These correspondences should be read as consistencies rather than demonstrated causal links.
A further, less anticipated finding is that outlet differences extend to genre. The three outlets did not merely emphasise different frames; they covered the topic in different journalistic forms: Asharq Al-Awsat through explanatory features, Al-Nahar through news reporting, and Al-Arab through opinion pieces. Because genre conditions the availability of frames (opinion writing invites evaluation; features invite contextualization), this generic divergence is itself a framing choice at the level of news production routines, of the kind Gamson, Gitlin, and Tuchman identified when they located frames in the organisational practices of journalism rather than solely in texts.
5.3. Limitations and Future Research
Several limitations qualify these conclusions. The corpus, while spanning three major outlets, comprises 53 items, and the outlet-level comparisons rest on modest cell sizes; the differences reported here are therefore descriptive patterns of emphasis rather than statistically confirmed contrasts, and replication with larger corpora, ideally supplemented by inferential tests such as chi-square where cell counts permit, is warranted. The analysis measures within-theme proportions, which capture the internal structure of each thematic field but not the relative prominence of themes; future work could weight themes by their absolute frequency and by the dominant-frame flag in the codebook. The study examines three dailies and cannot speak to broadcast, digital-native, or national outlets, where the uneven adoption landscape described by Harb and Arafat (2024) may produce different framing configurations. Finally, the design addresses frame building in texts, not frame setting in audiences; whether the disinformation-centred, state-remedy narrative documented here shapes Arab publics’ expectations of journalism’s future is an open question that framing-effects research could pursue. Extensions might also trace framing over time as generative AI matures, compare pan-Arab with national Arab presses, and combine manual coding with the computational approaches employed in recent cross-national studies.
6. Conclusion
Pan-Arab press coverage of journalism’s future in the era of artificial intelligence is structured around a coherent master narrative: a collaborative human-machine future, primarily threatened by synthetic disinformation and necessitating state-led legal intervention. The three leading dailies, however, interpret this narrative through their distinct editorial identities: Al-Arab emphasises threat and survival, Al-Nahar focuses on renewal and rights, and Asharq Al-Awsat highlights managed adaptation. By demonstrating both a shared repertoire and systematic divergence, this study addresses its two research questions and, in a context where such inquiry has been rare, affirms framing theory’s central insight that the future of journalism is not merely reported but actively constructed.