Media outlets face a challenge that goes beyond simply choosing between adopting or rejecting Artificial Intelligence (AI). Outlets are faced with designing a system that integrates digital infrastructure with professional standards. Qatar appears well-positioned for this task, thanks to its digital infrastructure, research investments, and the regional and international presence of its media.
Qatar launched its National Artificial Intelligence Strategy five years ago as part of a digitalization drive in various sectors—notably public service delivery. Given their role as a bridge between service providers and the public—especially in the case of the local press—the media were inevitably involved. The Qatar News Agency (QNA) has incorporated AI into various editorial processes, multilingual translation, and the production of video and visual content. It has also developed interactive digital services to help users access news, launching a virtual assistant on its website and app.
In 2024, QNA announced it would expand its use of these programs, in collaboration with Microsoft, reflecting a recognition of the transformations sweeping Qatari media. Yet it also challenged media outlets to develop professional guidelines to keep pace with rapid technological changes.
AI can help free journalists from routine tasks, granting them more time for verification, analysis and fieldwork. Yet this also carries risks. Without proper safeguards, speed can turn from an asset into a liability. Instantly published misinformation can be far more damaging than a few minutes’ delay in publishing the facts, because whatever reaches the audience first can become entrenched in their minds before any correction can catch up.
AI tools can speed up day-to-day tasks that once consumed hours in the journalistic workflow: transcribing audio interviews, monitoring audiences and trends on social media, reviewing thousands of documents to identify newsworthy information and translating content from global agencies, broadcasters and international media centers.
Yet the value of speed depends on accuracy. Most generative AI models operate on probabilities—predicting the most suitable word for a given context—without verifying the accuracy of the resulting information. This produces “hallucinations”: information that is coherently phrased yet factually incorrect, such as misattributed quotes or baseless statistics.
The challenge is compounded when dealing with the Arabic language. The volume of Arabic digital content available for training models is dwarfed by what is available in English by more than 60 percent. Furthermore, the diversity of Arabic dialects, coupled with variations in writing styles and rendering of translations, increases the likelihood of errors that may appear minor in terms of style, but can undermine the credibility of the media outlets publishing them.
Conversely, these challenges have presented an opportunity for research institutions to contribute to the development of models better equipped to understand Arabic and its local contexts. The Qatar Computing Research Institute (QCRI) has developed specialized Arabic AI tools, which are now used by the BBC, Al Jazeera and others.
Global events hosted by Qatar and elsewhere have tested the media’s AI capabilities and the rapid spread of misinformation. A prime example was the 2022 World Cup in Qatar, which was the target of systematic campaigns and an unprecedented wave of misinformation across various languages. This raises critical questions: Where did this (mis)information originate? Who bears responsibility?
Confronting Deepfakes
The risks go beyond the speed at which information spreads. Deepfakes target the very senses the public uses as evidence of an event: sight and hearing. AI-generated audio and video can give pure lies with the appearance of firsthand testimony.
Qatar offers a sobering lesson here. In 2017, hackers hijacked the QNA website and published fabricated statements, igniting a major diplomatic crisis. This demonstrated the ability of fake news to do damage before it can be refuted. If this was the impact of written misinformation, then faked audio-visual material could prove far more dangerous and persuasive, with political, economic and security repercussions.
Experts point to the impact of rapidly developing AI on the behavior of digital audiences, who are drawn to content that captures attention and drives engagement, rather than being dedicated solely to conveying facts. This presents media organizations with a growing challenge: “zero-click searches,” whereby users rely exclusively on summaries provided by search engines and AI tools, without visiting news sites to read the details.
Newsrooms must therefore find ways to detect fake or altered images and videos by reviewing files’ metadata, tracing their origins and cross-referencing them with independent sources.
Added to this is the issue of AI bias. AI models inherit biases from their training data regarding issues and groups, reproducing them in a superficially neutral format.
The U.S. National Institute of Standards and Technology (NIST) has warned that “a model trained on biased and erroneous data may lead to biased and inaccurate predictions.” This is critical for outlets that address a diverse, multinational audience and cover sensitive regional issues, where meaning hinges on specific choices of words.
One useful resource here is the Coalition for Content Provenance and Authenticity (C2PA) initiative—an open technical standard for verifying digital content by attaching authenticated information on how an image or video was created or modified.
When errors do occur, professional accountability is paramount. Media outlets remain legally and publicly responsible for everything they publish, whether written by an editor or generated by a model—and the tool cannot be cited to justify the error.
This entails three obligations. First is disclosure: readers have the right to know when such tools are used, particularly in producing AI-generated text, images and graphics. Second is human oversight: AI-generated content should not be published without a review by a responsible editor who knows the subject.
Third is protecting information. Inputting confidential documents or information from sources into external commercial platforms jeopardizes confidentiality which is critical to trust between journalists and sources. Then comes the issue of intellectual property; some AI draws on published journalistic material without the owners’ permission, necessitating caution when using or republishing its output.
From Tool Usage to Governance
Journalists do not need to race to adopt AI. Rather, outlets need to establish clear governance for its use. This demands a systematic institutional effort, along five tracks. First comes a written editorial policy defining which tasks may be assigned to these tools—such as transcription, preliminary translation and summarizing documents—and those that remain the exclusive domain of the journalist, such as assessing news value, evaluating sources and formulating allegations.
The second track focuses on training. Newsrooms require journalists who understand how to use these models. The Al Jazeera Media Institute, Qatar University’s College of Media and Northwestern University in Qatar are all working to develop skills to integrate AI into media practice.
A third track involves specialized fact-checking units that combine investigative skills with the precision of digital forensic analysis. The fourth entails partnerships with universities and research centers to develop tools that respect the linguistic and cultural nuances of Arabic and the Gulf, while testing their accuracy against clear standards.
The fifth and final track is about ensuring transparency with the public and fostering media literacy. One cannot combat fabricated content through technical tools alone. The better readers understand the digital content production process and the available methods for verifying sources, the better they can resist disinformation, in an environment of competing narratives in the region and the recent campaigns against Qatar.
The Paper Crisis and the Credibility Challenge
The adoption of AI comes as print journalism faces economic pressures and declining circulation. Many media organizations are shifting toward digital platforms in a battle for survival, chiming with research on the future viability of print journalism.
One example is Saudi Arabia’s Al Riyadh newspaper, which shuttered its presses in August. This does not necessarily curtail its reach, provided it can successfully channel its resources toward digital content. The core value of an outlet lies in the trust it has built with readers and public-interest institutions. Print journalism may need to embrace a broader digital presence to enhance its influence. Yet digital transformation does not entail abandoning the perennial function of journalism: verifying information and placing it in context.
The future of journalism will be shaped not by technology alone, but by the rules governing its use. If outlets can successfully combine innovation with verification and transparency, AI could bolster their presence, reach and analytical capabilities. Yet if speed is prioritized over credibility, these tools could erode decades of hard-won trust.
