AI Impact on Media and Journalism: The Industry That's Eating Itself

Media has been in crisis for 20 years. Print advertising collapsed. Digital advertising got eaten by Facebook and Google. Subscription models sort of worked for some outlets. Freelance rates went backwards. Newsrooms got smaller.

And then AI arrived, and it got worse.

The media industry's relationship with AI is uniquely self-destructive. Media organisations are simultaneously the victims of AI (their content was used to train AI models without meaningful compensation), the users of AI (they're deploying it to cut costs in their own newsrooms), and the reporters of AI (writing about how AI affects other industries while their own colleagues get laid off).

It's like watching someone write their own redundancy notice.

What's happening in newsrooms

Commodity content is being automated. Sports results, financial earnings reports, weather summaries, traffic updates, basic event coverage. AI generates this content now. Several major news organisations have been using AI for this for years. The difference is that it used to be supplementary. Now it's replacing the humans who used to do it.

First drafts of news stories. AI can take a press release, a set of facts, or a data source and produce a serviceable news article. It won't win any awards. It won't ask the awkward question at the press conference. But it produces readable content at a fraction of the cost. For news organisations under financial pressure (which is all of them), the temptation is overwhelming.

Sub-editing and copy editing. AI checks grammar, style, facts, and consistency. The sub-editors' desk, already reduced in most newsrooms, is shrinking further. Some online publications have eliminated human sub-editing entirely and rely on AI plus a final human check. The quality has... varied.

Content repurposing. Taking a long-form article and turning it into social media posts, newsletter summaries, video scripts, and podcast outlines. AI does this in minutes. The content teams that used to do it manually are being reduced.

Transcription and production support. Interviews transcribed by AI. Video footage logged by AI. Production schedules optimised by AI. The support staff around content production are being automated.

What's not being automated

Original reporting. Going to places. Talking to people. Asking questions that people don't want to answer. Building sources over years. Investigating stories that powerful people would prefer stayed hidden. This is the heart of journalism and AI cannot do it. AI has no sources. It can't knock on doors. It can't sit in a courtroom for six weeks. It can't build the trust that gets someone to share a document they shouldn't.

Opinion and analysis from trusted voices. Readers follow specific columnists and commentators because of their perspective, their voice, their track record. AI can generate opinions. But nobody cares about the opinions of a language model. They care about the opinions of people they trust.

Deep subject matter expertise. The defence correspondent who's been covering military affairs for 15 years. The health editor who understands the NHS inside out. The business journalist who can read a set of accounts and spot what the company is hiding. This expertise takes years to build and AI doesn't have it.

Visual journalism and documentary. Photojournalists, documentary makers, video journalists who tell stories through images. The physical, creative, relationship-driven nature of this work is resistant to AI.

The freelance massacre

If you're a freelance journalist or content creator, the AI impact is brutal and it's already happened.

Rates for commodity content have collapsed. Why would an editor pay a freelancer to write something AI can produce for free? Commissioning budgets have been cut because AI handles more of the content that freelancers used to provide.

The freelancers who are surviving are the ones who do things AI can't: original reporting, specialist knowledge, unique access, distinctive voice. The ones who produced serviceable-but-generic content are being priced out by AI.

This isn't a future prediction. It's the present. i know freelance journalists who've seen their income drop 40-60% in the last two years. That's not a slowdown. That's an industry restructuring beneath their feet.

The magazine and digital media picture

Magazines and digital media companies are using AI to produce content at scale. Some are open about it. Others are less so. The number of human writers at content-driven websites has plummeted.

The ones that were essentially content farms, producing SEO-optimised articles to capture search traffic, were always vulnerable. AI content farms can produce the same material at a fraction of the cost. The humans who wrote "10 Best Vacuum Cleaners for Pet Owners" for a living are in serious trouble. And honestly, as a reader, i can't say the quality difference is enormous.

But quality publications that depend on editorial voice, original reporting, and reader trust are also affected. Not because AI replaces their journalism, but because AI-generated content flooding the market devalues all content. When there's unlimited mediocre content available for free, even excellent content struggles to command a price.

What's growing in media

AI-specific journalism. Ironically, one of the growing beats in journalism is covering AI itself. The companies, the policies, the societal implications. Journalists who understand AI deeply enough to report on it critically are in demand.

Investigative and accountability journalism. Foundations, nonprofits, and some publishers are investing in investigative work because it's the one thing that clearly justifies human journalism. The model is changing (more foundation-funded, less advertising-funded) but the demand is there.

Multimedia and experiential storytelling. Podcasts, newsletters, events, community-building. The media businesses that are working aren't just producing content. They're building relationships with audiences. AI can produce content. It can't host a live event or build a community.

Communications and corporate content. This is the pragmatic option. Many journalists are moving into communications, PR, and corporate content roles where their writing skills, editorial judgement, and ability to tell stories are valued. The pay is usually better. The existential satisfaction is usually worse. Welcome to trade-offs.

What to do if you work in media

If you're a staff journalist: get closer to original reporting and specialist knowledge. The generalist news reporter is the most vulnerable role. The specialist with sources and expertise is the least. Build a beat. Build sources. Build a reputation for knowing things that AI doesn't.

If you're a freelancer: specialise or diversify. Having a clear specialism with unique expertise commands higher rates. Alternatively, diversifying into adjacent areas (copywriting that AI can't do, corporate communications, consulting) reduces your dependency on editorial commissioning.

If you're in production or support: the roles that involve judgement (editorial judgement, ethical decisions, strategic choices) are safer than the roles that involve processing (transcription, formatting, scheduling). Move towards judgement wherever possible.

If you're early career: learn AI tools deeply. Not to replace your journalism skills but to augment them. A journalist who can use AI for data analysis, research, and production is more valuable than one who can't. And understand that the business model of media is changing. The jobs that exist in five years might look very different from the ones that exist today.

The media industry has survived print's decline, the internet, social media, and a dozen other existential threats. It'll survive AI too. But "the industry survives" and "my job survives" are different statements. Know the difference.

The one thing to do today: look at the content you produce and ask yourself what percentage of it AI could produce to an acceptable standard. The percentage it can't is your competitive advantage. Make sure your employer knows it too.