March’s digital PR trends: Newswires are barely cited by AI search, and has AI-generated content finally matched human writing?

Newswires and syndications are barely cited by AI search

A new BuzzStream report analysed four million AI citations to uncover the types of news coverage influencing the AI results. And the findings look promising for digital PR.

Overall, the research found that 14% of all citations were from news publications. And of these news citations, 81% were from editorial coverage. Not affiliates or review-driven content – genuine, editorial news. Which is ideal for us, obviously.

But what’s also interesting is that press releases published through syndication only accounted for 0.32% of news citations and 0.04% of the entire dataset. Similarly, newswires only made up 0.21% of the full dataset.

So in other words, press releases distributed through newswires and syndication are barely cited by AI search. Which might come as a bit of a shock to press release distributors. Especially as more and more of them using AI visibility as a key selling point.

What does this mean for digital PR?

As an industry, we all know that newswires won’t drive anywhere near the value of true editorial coverage. And if your PR strategy is reliant on simply paying to appear on newswires in 2026, I’d argue that you’re not actually doing PR. Not very well at least.

So it’s good news that press release distribution services don’t really impact AI visibility. It means that brands relying on newswires for AI visibility probably won’t get the results they need – so will end up coming our way. Brilliant.

Syndication is a tricker one. It can be incredibly valuable. If the Reach journalist you’re pitching to syndicates your piece across all the relevant titles they work on, you could massively increase the impact of your campaign.

But this study seems to have focused on the newswire-y type syndications, which won’t be the bedrock of any self-respecting PR’s strategy. So ultimately, if they’re not getting cited – it’s more good news.

Is AI content now equal to human-written copy? New claim creates debate

This month Ryan Law, Director of Content Marketing at Ahrefs, caused a bit of a stir on LinkedIn. His blog titled ‘AI Content Wasn’t Good Enough. Now It Is.’, claims that AI content has improved to the point where it can (when properly directed) replace human content.

Check it out below. The blog (and the comment section) are well worth a read.

What does this mean for digital PR?

Law’s closing point stood out: “Until recently, AI content wasn’t good enough. Now, it is. The sooner we can admit that, the more time we have to focus on the parts of marketing where humans will have a longer, happier tenure.”

I think this is really, really interesting. Because if we’re looking at content from a purely “functional” point of view, I think he’s probably right. With a lot of set-up and some complex AI workflows, the content would probably do a good job for you. Particularly if the AI is focused on tasks like editing and updating existing pieces.

But to me, content should be more than “functional”. Whether you’re in PR, SEO or any other form of marketing – quality content is less about writing and more about thinking. With the right thought process behind it, you can influence your audience, engage a journalist, add nuance, project personality, showcase expertise, offer perspectives. And can AI do this? I’m not sure it can.

Yes, maybe you could spend hours inputting an enormous amount of background context, opinions, perspectives and examples into AI. Basically ‘downloading’ your brain. But you’d have to spend so much time and energy building (and then continually updating) this, I just don’t see how it could ever be worth the investment.

Another point to note is that as AI content creation improves, so will AI content detection. We know how brutal Google’s algorithms can be. Just look at the impacts of the site reputation abuse policy or its recent demotion of self-promotional listicles (covered in Feb’s column). So imagine if Google suddenly announced tomorrow that all content must be written by humans? Admittedly pretty unlikely, but you never know. If we reach a point where AI search is only pulling from AI-generated content, it could happen.

If it does, some of the highest quality AI-written pieces might survive (where marketers have heavily invested in ‘downloading’ their brains). But a LOT wouldn’t. And a LOT of brands would suffer in search. Especially if AI starts being used to the frequency that Law suggests.

So in short, while there’s definitely truth in what he’s saying, I don’t really agree with him. To me, human-written content is sticking around for the long run. But this is such an interesting discussion. And one that we need to continue having. Because AI, and all the problems and solutions it brings, isn’t going anywhere.

What else is new in digital PR & SEO?

• Wikipedia bans AI-generated content. Great news. Wikipedia’s such a key source of information online. It simply can’t afford hallucinations and inaccuracies. Will this push more publishers and platforms to ban AI-generated content? Let’s see. I really hope so.

• OpenAI pivots away from Instant Checkout within ChatGPT. The e-commerce feature launched in September hasn’t been a success. Walmart even reported that it resulted in 3x worse conversions… But don’t write off agentic shopping just yet, there’s definitely more to come.

• Google updates AI Mode to encourage clicks to recipe bloggers. This might be a small change, but I think it’s really positive. Google’s actually listened to feedback and changed something for the better. Fingers crossed this continues.

• AI citations ≠ AI results. Rand Fishkin recently covered the topic on LinkedIn. Much of the research around AI search frames citations as the sole creator of AI responses. In other words, each AI result is formed by the sources that have been cited (e.g. the links in a Google AI Overview). But this isn’t true. The reality is a lot more complex, with the sources sometimes even added afterwards through something called ‘post-hoc rationalisation’. It all gets very technical. But for me, the key takeaways are that; 1) the sources that AI cites aren’t always the ones responsible for the response, and 2) despite this, securing coverage in these citations remains one of the best approaches we have currently available for building AI visibility.