The jobs AI is replacing were already killing journalism. The real question is what we choose to build in their place.
The Issue
Imagine your job, as a journalist in 2026, is to write fifteen SEO articles a day about “best toasters 2026”. You earn £23,000. You have a journalism degree. And this week, an algorithm took your job. Is that a tragedy, or a symptom of a system already broken? Since when did churning out toaster rankings count as journalism at all?
AI didn’t create this crisis. It just held up a mirror to the content farm model that news organisations built themselves – and called it journalism. The jobs AI is taking were already stolen from journalists long before ChatGPT. The content farm model had systemically degraded what it meant to work in news. Low-paid freelancers and SEO strategies were used to generate large quantities of generic articles – as academic Piet Bakker documented as far back as 2012 – which is a far cry from traditional outlets that typically employ professional journalists.
The real threat to journalism was invented in a boardroom that decided clicks mattered more than truth. The advertising-driven model had already done the damage – incentivising volume over quality and turning journalists into content machines. Hedge funds like Alden Global Capital aggressively acquired local newsrooms across the United States and gutted them – extracting profit while communities lost their watchdogs. In the UK, Reach, which owns the Daily Mirror and hundreds of local titles, pursued the same playbook. The message was clear: quality journalism was already losing before AI entered the room.
Current responses to AI in journalism are treating the symptom rather than the cause. Notable news organisations Reuters and the BBC have both introduced internal AI guidelines. The current EU AI Act also includes transparency requirements for labelling AI-generated content. At the other end of the spectrum, some outlets have chosen to ban AI content altogether. The very outlet publishing this piece has been publicly cautious about AI-generated content – stating that GenAI tools are exciting but are currently unreliable. However, regional and local titles in the UK are among the most vulnerable financially. Reach, which owns the Daily Mirror, Express, and Star, announced 450 redundancies in a single year.
Whether newsrooms choose to label AI or ban it entirely, both responses share the same blind spot – these guidelines tell readers what but not why. Neither addresses the broken economic model driving AI adoption in the first place.
Recommendations
The toaster article journalist deserves better. But so does journalism itself. Fixing it requires dismantling the economic model that broke it in the first place. Firstly, the government should establish a public funding model for local news, similar to the Nordic model. It has already proven results: research by the Australia Institute has shown that the model helped support media diversity and public interest journalism. It addresses market failure directly rather than relying on broken ad revenue.
Secondly, AI companies that train their models on journalistic content should be legally required to fund journalism in return. This would allow news organisations and governments to redirect money from those profiting from journalism back into producing it. There is already working evidence: the Australian government has introduced new laws that could force tech giants such as Meta and Google to compensate local news outlets for publishing their content. Known as the News Bargaining Incentive, these draft laws encourage platforms to enter deals with news publishers as the preferred new model. Furthermore, research done by Nadel (2023) recommends an antitrust exemption that allow news services to negotiate copyright fees from Facebook, YouTube, and other large online platforms. The preexisting success of the News Media Bargaining Code in Australia shows that this is legally and politically achievable.
Lastly, journalists need stronger union protection. Journalists need agency in how AI is deployed rather than having it imposed on them. This matters because the content farm model thrived precisely by keeping journalists atomised, underpaid, and unable to organise collectively. If we want journalism to emerge from this crisis in better shape than it entered, workers need a seat at the table when AI decisions are made.
Conclusion
The panic surrounding AI in journalism is understandable – but it is misdirected. These disappearing jobs are the product of a broken economic model that exploited journalists’ professional idealism, degraded editorial standards, and called the result journalism. AI did not build the content farm. Boardrooms did.
This is not an argument for complacency. It is an argument for ambition. The disruption AI brings to journalism is real – but so is the opportunity it creates to finally demand something better. Public funding for local news. A levy on the tech giants that have gorged themselves on journalistic content for decades. Stronger protections for the journalists who remain.
To put it simply, we need to save journalism. And that means building a system where it is not about writing fifteen toaster articles daily – but about holding power to account, serving communities, and doing the actual democratic work.
