Will artificial intelligence be the death of journalism? - Index on Censorship

In mid-May 2023, The Irish Times published an article that accused women who use fake tan of mocking those with naturally dark skin. The op-ed was initially said to be written by Adriana Acosta-Cortez, a 29-year-old Ecuadorian health worker living in north Dublin.

But no such person existed. “The article and the accompanying byline photo may have been produced, at least in part, using generative AI technology,” read an editorial in The Irish Times four days after the piece first published.

Two months later, HoldtheFrontPage— a news website for journalists with a focus on regional media across the UK— published an investigative piece documenting how artificial intelligence (AI) was used to launch a publication purporting to be called The Bournemouth Observer, which turned out to be a fake newspaper. “It was obvious that the content was written by AI because the writing was so bad,” editor of HoldtheFrontPage, Paul Linford, told Index. “But since then, AI has got much better at writing stories, and I suspect it will eventually become harder to spot when writing is being done by AI or real journalists,” said Linford.

Index on Censorship was also caught out by a journalist calling themselves Margaux Blanchard whose article was published in the Spring edition of the magazine. Ironically it was about journalism in Guatemala and written by AI.  Others – Wired and Business Insider – also fell victim to “Margaux”.

James Barrat claimed AI “will eventually bring about the death of writing as we know it.” The American documentary maker and author has been researching and writing about AI for more than a decade. His previous books include Our Final Invention: Artificial Intelligence and the End of the Human Era (2013), which ChatGPT recently ingested. “There are presently ongoing lawsuits about this because OpenAI took my book [without my permission] and didn’t pay me,” Barrat explained. “Right now, if you tell ChatGPT ‘write in the style of James Barrat’ it doesn’t produce an exact replica, but it’s adequate, and machine writing is getting better all the time.”

AI “will eliminate 30% of all jobs”

In early September, Barrat published The Intelligence Explosion: When AI Beats Humans at Everything (2025). The book makes two bold predictions. First, AI has the potential in the not-too-distant future to potentially match, and perhaps even surpass, our species’ intelligence. Second, by 2030 AI will eliminate 30 percent of all jobs done by humans, including writers. Freelance journalists will benefit in the short term, Barrat claimed. “Soon a basic features writer, using AI, will be able to produce twice as much content and get paid twice as much,” he said. “But in long run the news organisations will get rid of [most] writers because people won’t care if content is written by AI or not.”

Tobias Rose-Stockwell did not share that view. “There will always be a market for verified accurate information, which requires humans,” the American writer, designer, technologist and media researcher said. “So truthful journalism isn’t going away, but it’s going to be disrupted by AI, which can now generate content in real time. This will lead to more viral falsehoods, confusion and chaos in our information ecosystem.”

Rose-Stockwell elaborated on this topic in Outrage Machine (2023). The book documents how the rise of social media in the mid-2000s was made possible by algorithms, which are mathematical instructions that process data to produce specific outcomes. In the early days of social media users viewed their feeds in chronological order. Eventually, though, Facebook, Instagram, X, TikTok, and other social media platforms realised it was more profitable to organise that information via algorithmic feeds, powered by artificial intelligence and, in particular, machine learning (ML) where AI is used to identify behaviours and patterns that may be missed by human observers. ML tools analyse users’ behaviour, preferences, and interactions, keeping them emotionally engaged for longer. “Feed algorithms are much better at re-coordinating content than any human ever could,” said Rose-Stockwell. “They can even create bespoke little newspapers or television shows for us.”

“AI is already in the process of rapidly transforming journalism,” said Dr Tomasz Hollanek, a technology ethics specialist at the University of Cambridge with expertise in intercultural AI ethics and ethical human-AI interaction design. “As AI systems become more adept at producing content that appears authentic, detecting fabricated material will get harder.”

Hollanek spoke about editors giving journalists clear guidelines about when and where AI can be used. The Associated Press, for instance, currently allows staff to test AI tools but bans publishing AI-generated text directly.

“What’s important about these guidelines is that while they recognise AI as a new tool, they also stress that journalism already has mechanisms for accountability,” said Hollanek. He also criticised the sensationalist tone journalists typically take when writing about AI, pointing to unnecessary hype, which leads to distorted public understanding and skewed policy debates.

“Journalists strengthening their own critical AI literacy will make the public more informed about AI and more capable of shaping its trajectory.”

AI’s role in news “needs to be trackable”

Petra Molnar, a Canadian lawyer and anthropologist who specialises in migration and human rights, claimed “the general public needs to understand that AI is not some abstract tool out there, but it’s already structuring our everyday lives.”

Molnar said there is an urgent need for public awareness campaigns that make AI’s role in news and politics visible and trackable. She described companies such as Meta, X, Amazon, and OpenAI as “global gatekeepers [of information] with the power to amplify some voices while silencing others, often reinforcing existing inequalities.”

“Most people experience AI through tools like news feeds, predictive texts, or search engines, yet many do not realise how profoundly AI shapes what they see and think,” said Molnar, who is the associate director of the Refugee Law Lab at York University, Toronto— which undertakes research and advocacy about legal analytics, artificial intelligence, and new border control technologies that impact refugees.“AI is often presented as a neutral tool, but the reality is that it encodes power.”

Last year, Molnar published The Walls Have Eyes: Surviving Migration in the Age of Artificial Intelligence (2024). The book draws attention to a recent proliferation across the globe of digital technologies that are used to track and survey refugees, political dissidents, and frontline activists crossing borders in times of conflict. Molnar claimed that “AI threatens to accelerate the collapse of journalism by privileging speed and engagement over accuracy and depth.” She cited examples of journalists using OpenAI-generated text tools to churn our surface-level articles that echo sensational framings around migration, without investigative depth.

“Automated systems may generate content that looks like journalism, but it’s stripped of accountability and critical inquiry that’s required to tell complex stories,” said Molnar. “Journalism’s future depends on human reporters who can investigate power and rigorously fact check, something AI simply cannot replicate.”

Human oversight

Sam Taylor, campaigns & communications officer at the UK’s National Union of Journalists (NUJ), shared that view. “Editors and writers should exercise caution before using AI in their work,” he said. “Generative AI often draws on databases from the internet that contains stereotypes, biases, and misinformation.”

“To maintain and strengthen public trust in journalism, AI must only be used as an assistive tool with human oversight,” said NUJ general secretary, Laura Davison.

Everyone Index spoke to agreed that AI, for all its flaws, offers journalists enormous benefits, including automating mundane routine tasks, like transcription and data sorting. AI can also make data journalism, exploring large data sets to uncover stories, much more accessible as AI can crunch the data and identify interesting nuggets far faster than a person can. This will leave journalists with more time and energy for critical thinking, and ultimately, to tell more complex and nuanced stories.

But there was also an overwhelming consensus that AI cannot fact-check accurately or be trusted as a credible verifier of information. Not least because it suffers from hallucinations. “This means due to the complexity of what is going on inside it hallucinates,” James Barrat explained. “When this happens, AI gets confused and tells lies.”

“The jury is still out on whether or not this hallucination problem can be solved or not,” said Tobias Rose-Stockwell. “Journalism must remain grounded in ethical responsibility and context,” said Petra Molnar. “What we need is human judgement, applied critically and ethically, supported by but not replaced by technology.”

Is AI a threat?

Anyone who believes in journalism’s primary mission, to challenge power by investigating the truth, is undoubtedly likely to agree. But is this wishful thinking from a bygone era? James Barrat believes so. He points out that, eventually, we may not have the option to choose. “A scenario that could happen in near future is that AI could become hostile to us,” he said. “AI could take control of our water and our electrical systems. Just recently, a large language model (LLM) agreed that its creator should be killed.”

Barrat mentions an interview he did with the British science fiction writer and futurist Sir Arthur C Clarke, before his death, aged 90, in 2008. Clarke co-wrote the screenplay for 2001: A Space Odyssey (1968). Directed by Stanley Kubrick, the Oscar winning film tells the story of an AI-powered computer— aboard the Discovery One spacecraft bound for the planet Jupiter— called HAL. Eventually, HAL experiences a program conflict, malfunctions, and, to defend itself, turns on its human crew members.

Arthur C Clarke told Barrat, “Humans steer the future because we are the most intelligent species on the planet, but when we share the planet with something smarter than us, they will steer the future.”

“AI’s intelligence grows exponentially,” Barrat concludes. “As it gets smarter, we will stop understanding it. We really are inviting disaster.”

The autumn issue of Index magazine, titled Truth, trust and tricksters and published on 26 September, looks at the threats that artificial intelligence poses to freedom of expression