By Xhabir Deralla

As someone who entered journalism under a communist regime in which a journalist was legally defined as a “socio-political worker,” I have learned to be cautious whenever rigid limitations are imposed on journalistic work. Journalism cannot breathe under prescriptions that treat the profession as something to be controlled rather than strengthened through freedom, responsibility and professional integrity. In journalism, as in all other spheres of human society, openness, inclusiveness, respect and progress are among the values I have always cherished most. That remains true almost four decades later, as I enter discussions – and write books – on artificial intelligence: its history, its meaning as a phenomenon, and its use in public life, media and journalism.

The way I see it, AI is not simply a technological novelty. In the context of the media and news industry, it is a test of whether journalism still has the courage to adapt and embrace this new stage of evolution. Newsrooms have always changed with technology: the printing press, the telegraph, radio, television, satellite broadcasting, the internet, social media and smartphones. AI is another turning point — but this time, the challenge is not only production speed. It is also about research, verification, monitoring, multilingual access, disinformation analysis, accessibility, audience understanding and the survival of public-interest journalism under pressure.

Let me say this at the very beginning, so there are no misunderstandings about where I stand. I am perfectly aware that the future of journalism will not belong to media that blindly use AI. But it certainly will not belong to those that theatrically reject it either. The latter are already choosing obsolescence. In my view, the future belongs to newsrooms that know how to combine human judgment with technological capacity, at least for now, and in the foreseeable future.

Media cannot demand democratic modernization while remaining technologically and professionally stagnant. In a rapidly changing world, authoritarian powers and anti-democratic actors are already using technology to influence societies through sophisticated propaganda and hybrid operations — targeting politics, defence, the economy, culture and public life. Artificial intelligence is increasingly embedded in these operations, often in entirely non-transparent ways. Journalism and the media must adapt and pick up the pace. Urgently.

Yet much of the debate about AI in journalism remains trapped between rigid refusal, blanket discrediting of any AI use, and defensive, old-fashioned caution. In the meantime, those who exploit the media ecosystem for anti-democratic purposes continue to operate smoothly under the democratic radar.

Global newsroom transformation: The positive contribution

While AI is already generating entire websites and vast amounts of content across the world – fueling debates, anxieties and predictions that journalism is approaching its end – serious newsrooms are responding differently. They are actively exploring how AI can be integrated responsibly, efficiently and transparently, not as a substitute for journalism, but as a powerful new technology whose proper use can strengthen public-interest reporting.

To say that AI’s potential contribution to the public-interest side of journalism is significant would be an understatement. In research, fact-checking, automated monitoring and verification, its achievements are already historic, not only in terms of speed, but also in volume, depth and precision. AI can help journalists to process large volumes of documents, transcripts, public statements, datasets and social-media content. It can detect patterns in narratives, compare claims across multiple sources, identify manipulated or recycled content, and support multilingual monitoring of disinformation campaigns with a capacity that was unimaginable only a few years ago.

Some argue that this means truth is being outsourced, or even surrendered, to machines. I would argue the opposite. This is precisely one of the most responsible ways to engage with AI: keeping the human in the loop while expanding the capacity of journalism. Journalists use technology to process more, see more, compare more and detect more, while editors retain responsibility for judgment, context, verification and publication.

Studies on automated fact-checking repeatedly show both promise and limits. AI can assist verification workflows, but dynamic events, local realities and complex claims still require strong human oversight. That does not mean that journalism has surrendered truth to machines. It means that journalism has acquired new capacity, and must now prove that its responsibility is strong enough to use that capacity well.

The question, once again, is not whether machines will replace truth, but whether humans will use them in its service.

Let’s take a look at some positive examples.

The Associated Press is one of the clearest examples. AP has long used automation for structured, data-based reporting, including corporate earnings stories. It has also described the use of AI and automation in transcription, translation, information management, localization, shot lists and dataset-based reporting. The example of the Associated Press shows that AI can be integrated into newsroom workflows without replacing journalism itself.

Le Monde offers a strong European example of a balanced and cautious model. The French newspaper has adopted an AI charter stating that generative artificial intelligence cannot replace editorial teams. At the same time, Le Monde uses AI-supported translation, followed by human translation and editorial review, as well as AI-assisted audio generation and support in comment moderation. In other words, Le Monde uses AI in production, access and reader services, but not as a substitute for editorial authorship.

JournalismAI / LSE’s “Generating Change” research surveyed 105 news organizations from 46 countries. This helps us understand that the question of AI in journalism is not a “big tech” issue, nor a U.S.-only debate, but part of a global newsroom transformation.

The Reuters Institute’s Digital News Report 2025 provides the broader context for why journalism must adapt. The report describes traditional news media as struggling to connect with much of the public, with declining engagement, low trust and stagnating digital subscriptions. It also points to the continued shift toward social media, video platforms and fragmented alternative media environments. This reinforces the argument that media cannot claim to change the world while refusing to change themselves.

That is why media cannot afford to demand that societies modernize, democratize and adapt to reality while newsrooms themselves remain trapped in outdated professional reflexes. If journalism wants to remain a force for truth in a world of information overload, propaganda and synthetic manipulation, it must learn to use AI better than propagandists, political machines and commercial manipulators do. Failing to do so will not protect journalism from AI. It will only help journalism fade into oblivion.

The danger

Of course, there are dangers. Critics are right to say that AI can accelerate manipulation, synthetic content, plagiarism, lazy journalism, fabricated sources and dependence on opaque platforms. But these dangers did not begin with artificial intelligence. Was this not also true in the Gutenberg era, when the printing press made ideas, truth, lies, propaganda, science and hate travel faster than ever before? What is different today is the speed, scale, variety, personalization and persuasive power of the tools. The principles, however, remain the same. Human perception too.

Plagiarism existed before AI and will exist after it, assuming, of course, that anything is still allowed to exist “after AI” in the apocalyptic imagination of its loudest critics. Fabricated sources are not new to journalism. Neither are laziness, sensationalism, political obedience, commercial pressure, propaganda or the betrayal of public trust. AI can make all of this faster, cheaper and more convincing, but it does not invent the ethical failure. It amplifies it.

That is why the central question is not whether AI is dangerous in itself, but who uses it, for what purpose, under what rules, with what transparency, and under whose responsibility. The real ethical battlefield is human agency: the decisions of journalists, editors, media owners, political actors, technology companies and institutions. Machines can generate, classify, summarize, imitate and accelerate. But they do not carry conscience, professional integrity or democratic responsibility. At least in the foreseeable future, those remain human duties.

This is precisely why journalism must not respond to AI with fear or ritualistic rejection. It needs to respond with standards, literacy, accountability and editorial control. The answer to manipulation is not technological stagnation. The answer is better human judgment, supported by better tools and governed by clear professional and ethical responsibility.

AI does not abolish responsibility. One cannot say, “It wasn’t me, it was the AI.” On the contrary, the presence of AI exposes, in certain ways, whether responsibility existed in the newsroom in the first place.

The real standard

When we discuss the use of AI in newsrooms, or in work related to the monitoring of socio-political processes (elections, human rights, good governance, public policy, academic research or disinformation), the debate should not be reduced to a Shakespearean dilemma “to AI or not to AI.” The core issues remain human editorial responsibility, transparency, verification, source protection, privacy, copyright and accountability. As has always been the case.

Where these standards are absent or violated, AI is not the one to blame. Humans are. It is as simple as that.

Therefore, what media organizations, civil society and thought leaders need to discuss is how to protect the core standards of journalism in a time of authoritarian offensive, information disorder and technological acceleration. AI is now part of that reality. Not as a substitute for journalism, but as a new companion in the newsroom. Its contribution should be harmonized with clear ethical standards, transparency, verification and editorial accountability.

In a way, the debate recalls earlier technological shifts in journalism. The pen and notebook were joined, and in many cases replaced, by the dictaphone. Later came digital recorders, smartphones, search engines, databases, analytics instruments and social media monitoring. Journalists did not seriously debate whether a recorder with voice activation was, in itself, a threat to journalism. The real question was always whether the journalist used the technology professionally: to record accurately, protect the source, verify the statement, understand the context and report truthfully.

AI is incomparably more powerful, more complex and more consequential than a recorder. But the principle remains the same: technology – now intelligent technology – does not define journalism. Standards do. Human responsibility does. Editorial integrity does.

The task, therefore, is not to dramatize the presence of AI in journalism, but to define the conditions under which its use strengthens, rather than weakens, public-interest reporting. In an era of authoritarian influence operations and information disorder, journalism cannot afford technological innocence. It needs ethical clarity, professional courage and the capacity to use new technologies better than those who weaponize them against democracy.

Let me conclude with a personal note. Journalism must remain human in purpose, but it cannot remain analogue in method. This comes from someone who started with a UNIS Sarajevo typewriter at the end of the 1980s. When the first PCs entered our newsroom, some were suspicious and reluctant. Others embraced the leap forward. Here we are, far beyond Gutenberg, but still struggling with the same questions.

Journalism does not defend its integrity by refusing new technologies. It defends its integrity by using them better than propagandists, manipulators and political machines do.

That’s the story.

Author’s note on AI use:
Since the beginning of my engagement with generative AI, ChatGPT has been a natural companion in my writing, research and thinking process — not as a substitute for judgment, authorship or responsibility, but as support in research, language refinement, testing arguments, editorial advice and visual creation. The same applies to this article. The concept, position, final text, editorial choices and responsibility remain entirely mine.

YOUR SUPPORT TO CIVIL MATTERS.

Support independent journalism, democratic resilience, and the defense of freedom of expression. Help CIVIL counter disinformation, hybrid threats, and authoritarian influence.

Contribute NOW