Online News And Generative AI -- Barry Adams // Polemic Digital

The Voices of Search podcast is a proud member of the iHearEverything podcast network. Looking to launch or scale your podcast? iHearEverything delivers podcast production, growth, and monetization solutions that transform your words into profit. Ready to give your brand a voice? Then visit iHearEverything.com.

Bridge toll, California. Customer service number.

Highway miles to the gallon, Ford Focus.

Thailand cave rescue operation.

What is Schema F?

Best wine bars in San Carlos, California.

Best Western hotels.

How old is Ronaldo?

What happened with Big Brother?

What's a good engagement?

How long before a wedding should I send out?

How many days during the first series?

Use IMAP to check email on other email clients.

Identify fonts from where to find.

Welcome to the Voices of Search podcast, an I Hear Everything production. In this podcast, we'll share the news, knowledge, and strategies you need to to navigate the ever-changing world of SEO. Ready to expedite your company's organic growth efforts? Sit back, relax, and get ready for your daily dose of search engine optimization wisdom. Here's today's host of the Voices of Search podcast, Tyson Stockton.

Welcome to the Voices of Search podcast. My name is Tyson Stockton from Previsible.io, and today we're going to be discussing generative AI and SEO in news publishing. Joining us today is Barry Adams, who is a specialized SEO consultant for news publishers at Polemic Digital, which is a specialized SEO consultancy providing services to news publishers around the world. They've worked with many of the world's largest media brands as well as niche and local publishers, helping them grow through organic search traffic. All right. And with that, let's dive into the topic of generative AI and SEO for news publishing. Barry, welcome to the podcast.

Thank you very much for having me, Tyson.

Excellent. So, I mean, obviously generative AI has been a hot topic across the entire year. I'm interested to hear your take and kind of within this vertical of like news and publishing. So maybe to start off for the listeners, what unique challenges do you see within news and publishing, like in regards to generative AI?

Yeah, I mean, I think the general mood has changed a little bit since generative AI came onto the scene the first time around when it first appeared and these large language models started gaining traction. And, you know, all these, these promises were made and a lot of hype existed around it. I think some news publishers were afraid that generative AI might replace news to a certain extent or replace the art of journalism. And I think that turned out to be not true. If anything, I think that the mood is now changed to a more positive view where news organizations see a potential in generative AI to help them be better at their jobs and do better journalism rather than seeing it as an existential risk to journalism. I think initially the fear was that generative AI systems would be able to report the news and write articles about current events. In a similar or same way as a lot of news outlets are doing, you know, especially news outlets that don't necessarily have a lot of journalists on the ground in the trenches, so to speak, and that rely on third-party sources to report news on their own websites. I think they really felt a sort of a fear coming out of generative AI. But I think now that that initial shock has dissipated a bit, talking to my clients, I think they're now more looking at ways to explore generative AI in a positive way and finding ways to implement it as part of their workflow. To make journalists' lives easier and help them do their jobs better and allowing journalists to focus on what a journalist should be doing, which is, you know, investigating facts, covering stories, finding sources, and then writing up what they discover and make sure that it's fact-checked and well-presented before it gets published. And I think generative AI can definitely play a positive role there. And I don't feel that it's going to be a major threat to news as an industry, online news in particular.

Interesting. And I feel like there's a lot of parallels in that to the rest of the industry. It's like, of course, there's the initial kind of shock element of, like, job's going to be out. Are we going to be obsolete? And then it sounds like within news and media, it's more of the recognition of, hey, this is a tool that we can leverage for it. What are some of the better applications that you've seen as far as using this as a tool or as a means to improve a efficiency within publishing?

Yeah, I mean, some of the more interesting applications I see, or I hear some of my clients experimenting with, is, for example, generating headlines. It's one of the hardest things often of writing a news story is to generate a headline, a headline that captures the essence of the story, but also that is optimized for visibility in Google's search ecosystem, as well as the social media landscape where a lot of traffic to publishers can also come from. And it's not something that is intuitive for many journalists because it's not really something that's part of their skill set. They know being a journalist is not being an SEO and it's not being a social media manager. So with generative AI, you can basically write your article as you would in any journalistic context and then let the generative AI suggest a few headlines for you. And that actually can lead to really powerful, really potent headlines that capture the story really, really well, but are also optimized for maximum visibility in Google's news ecosystem and/or on Facebook or on other social media channels, on Reddit, you name it. So that's one of the more interesting aspects that I've seen generative AI being used for.

Interesting. And with that, are there any particular contextual elements that you've seen included in the prompts or anything that's like helps improve what that output or what those suggestions are? Like, I assume it's taking the written article, adding that in, and then asking for X amount of suggested titles. But have you found like any other contextual elements that are helpful in including in the prompts?

Yeah, I think if you just let it generate titles, you're not going to get the real power out of it. You have to sort of put a certain emphasis on things. For example, if you write an article about Donald Trump, you want to make sure the headline contains the word Trump in it because Google as a search engine still requires fairly explicit signposting in news headlines to understand what an article should be ranked for. So in that case, your prompt should be, give me 10 headlines that all have a focus on Donald Trump as the main topic of the story. And that sort of a prompt will be a lot more geared towards generating a successful headline for that article, especially in the context of the news ecosystem. Whereas on social media, and this is one of the interesting bits that a lot of publishers struggle with, because headlines on social media tend not to be as explicit or have to be as explicit as headlines on search. And they can be a bit more clickbaity and a bit more sensationalist. So you don't necessarily have to have the word Donald Trump in there. You know, the context of a news story on social media is different than in a search result. So you can then give a prompt and say, make this an engaging headline that will encourage people to click on it. And then you get very different sort of headlines based on the exact same article that you can then sort of pick and choose and say, okay, which one do we want to attach to this particular story?

Interesting. I like this application too, because I would, I would assume given the nature of news and media, maybe there are sometimes limitations on the timeframe that the systems are being trained on as far as, you know, what information they've already had and ingested. But in these cases, or in this case that you're explaining, it's like, you're providing the article, you're not relying on it to have necessarily already context to maybe the information of it. But have you ran into any potential limitations on time intervals of when these generative AI models are being kind of trained under?

Yeah, I mean, I think for OpenAI, the cutoff was somewhere in 2021. So anything that happened after 2021, and you want, say, for example, to write an explainer article with a bit of background on a current event, you will always have to double-check it because A, it's probably not entirely up to date, and B, you can't actually necessarily trust the output of an LLM, even if it is an event that has taken place way before 2021, because the LLMs are notorious for just making up stuff. So you will always need that human editorial oversight. But, you know, in that same context, if you want to write an explainer article, what you could do is ask the LLM to summarize a Wikipedia page for you and make that a bit more accessible rather than purely academic language. And that can be the foundation of a really good explainer article. And you can sort of pick and choose which elements then you want to highlight in your story, in your explainer article on the website. So it's about just going beyond just seeing it as a purely generative text experience and giving it a specific instruction or set of instructions based on either content you provide yourself or content that you feel is more reliable than the entirety of the web. So if you're quite picky with your choices, you can tell ChatGPT to say only use sources with like academic citations or things like that. But at every stage, the human element has to remain present. I think LLMs, no matter how powerful they are, and they are very, very powerful, They are not flawless. They are still machines and they can help you do your job better, whatever your job is in any form of content generation. But I do think there needs to be that human editorial oversight to do the fact-checking, to make sure it's accurate, to make sure the language used is correct as well and appropriate for that particular story. But if you have those guardrails in place, it definitely helps journalists and content creators all over the world be better at their jobs, be more efficient with their time, and spend less time doing the things that generally people don't enjoy doing.

Time for a 1-minute break to hear from our sponsor, Previsible. So you're looking for SEO help, and you got a couple of options. You could start replying to spam from agencies that claim they can get you to rank number 1 on Google. You can pay an hourly rate for a consultant who will inevitably nickel and dime you with hourly charges, or you can work with a cookie-cutter agency to quickly launch a strategy-less project with low success rate. None of those sound very good, now do they? Well, that's where Previsible's integrated consulting model comes in. Previsible draws from a collective 40 years of SEO and digital marketing experience to unlock your organic growth opportunities. They build custom solutions that combine strategy, technical expertise, content, and reporting to effectively operationalize SEO for your business. Previsible's 4-stage approach ensures that your SEO programs thrive by starting off with a strategy-first approach. Then they support you in your efforts to create quality content, help you identify technical issues, and most importantly, they'll work with your cross-functional teams to integrate your SEO strategies to make sure that your SEO budget actually drives results, not just your agency's bottom line. So join brands like Yelp, eBay, Canva, Atlassian, Square, all who rely on the SEO consultants at Previsible. For more information, go to previsible.io. That's Previsible, P-R-E-V-I-S-I-B-L-E dot I-O.

And it feels like in that description too, it's more of just being a little more deliberate in where time is being spent and using this And I would assume for probably a lot of the journalists and some of the more kind of tedious elements that aren't necessarily the aspects of their work or the aspects of the job that genuinely excite them, or kind of like your piece earlier about making for titles.

Yeah.

Not being necessarily a marketer is probably not their favorite part of the job. I mean, making some assumptions here, but it sounds like a lot of these applications too are more of areas that could be potential pain points for the journalists rather than like, oh, it's taking something away that I feel like I was offering significant value under.

Exactly. Exactly. Another example is something that actually predates large language models, but is a form of automation that a lot of publishers had already been doing, which is internal linking. A lot of news publishers, when you start writing an article in their CMS about, say, a political topic and you mention a political figure like Joe Biden or Kamala Harris, the CMS will either encourage you to put a link in there to the Joe Biden topic page or the Kamala Harris topic page, or will do it automatically when the article is being published. And those sorts of semi-AI systems have actually been part of publishing for quite a long time. And I think the current generation of AI tools has made those tools even better. Recognizing entities, recognizing relationship between entities, and recognizing when it is appropriate to put a link into an article to a topic page around an entity that you mentioned in that article. Article. And the automation has gotten to such a stage that it tends to do at least as good a job, if not a better job, than humans manually putting in links to topic and tag pages in the articles themselves. So more publishers now saying to their journalists, put a link in to a related article if you want, especially if it's like a follow-up story on an existing news event, but you don't have to worry about putting in links to topic pages and tag pages because our systems will take care of that for you. And that makes the workflow for the journalist a lot easier. It also makes the internal linking a lot more consistent for the publishing website because, you know, the AI systems pretty much do it all the time consistently across the whole website, across every article being published. And we know that that sort of internal linking to topic pages is an incredibly powerful tactic that helps build the SEO value of that publishing website over time. It increases topic authority. And it basically results in long-term growth for that website and long-term trust and authority signals in Google's news ecosystem. So that's one way where, you know, the evolution of AI and not necessarily specifically generative AI, but just these automated systems based on machine learning mechanisms, taking tasks away from humans that humans don't really want to do, that humans see as overhead, as a necessary evil. 'Cause like you said, marketing is not necessarily a nice word in journalism. You know, it seems sort of almost a dirty word because marketing in the context of news publishing tends to be related to commercialism or to advertising, to making money off of news. And journalists generally don't concern themselves too much with that. They just want to do proper journalism. So the more we can use these tools to remove these tasks from the day-to-day workflow and liberate journalists to focus on actual journalism, the better it is for the website, the better it is for the journalists in the trenches as well.

Interesting. I like to go back just a little bit. Interested in the component that you're saying about kind of internal linking and to see the potential huge value add in regards to that. Are there any particular guardrails that you'd seen or you'd recommend to the listeners when they're exploring using some of these models for efforts around internal linking?

Yeah. It is something you want to at least keep an eye on and not just let it run automatically without any oversight. It is something you want to do a bit of sanity checking on. First and foremost, the real power of internal linking to topic pages comes to the fore when it aligns with your journalistic expertise as a publisher. There's only an X amount of websites that are seen as general news publishers that can write about anything. That's maybe a handful per country, and most other news publishers tend to be more focused on specific niches and specific topics. So you want to make sure that when you do internal linking to your topic pages and your tag pages, it is to topics and tags that you as, as a publisher are confident that you can deliver value on, that you regularly write about, that you have good journalism on and good quality of unique content. If you just liberally link to every topic that is ever mentioned on your website, you will end up with hundreds of thousands of topic pages, a lot of which have maybe one or two articles associated with them. And that has the exact opposite effect. That's you signposting that you're a specialist in nothing. That's you basically telling Google, we write about everything and we're good at nothing. So you have to sort of pick your battles. You have to be confident about what you are good at or what you want to be seen to be good at as a publisher and focus your internal linking as well as, of course, all your optimization efforts and your content efforts on those topics and those areas that you want to dominate in Google Search and in Google News. And the second is that you have to be careful for different phrasings and different spellings. Some systems will automatically generate tag pages for you, and there's a real risk you end up with 3, 4, 5, 6 tag pages around basically the exact same topic. For example, Joe Biden, President Biden, President Joe Biden, Biden on its own. So you have to sort of do a bit of a sanity check as well and make sure that You know, you might use different phrasings in the article, but you have to have a unified topic page that is consistent around one phrasing. And then yes, the anchor text that points to that topic page can vary from article to article, but you're always talking about that one same entity, that one same topic. So you just want to have one topic page around that particular entity and not just start creating new ones and new phrasings. Plus, you know, journalists being human, They occasionally make spelling errors, and it's very easy to then have an automatic system create a tag based with an incorrect spelling. So again, there has to be a process in place where you don't just let these systems automatically create tags. And also, you shouldn't let journalists create tags themselves. There has to be an ownership, a human ownership of that process, and the management built in there as well to occasionally review tags and see whether you need to add new ones, Or you need to delete older out-of-date tag pages around content that maybe you're not covering anymore and have no expectation of ever covering again.

Interesting. I have several questions or follow-ups on that, but I think I'm going to save some of that for our conversation tomorrow. It's just given me kind of flashes of different websites I've worked with in the past that have had some of these substantial issues on, yeah, over-indexation played page bloat and everything else. But to circle back into the generative AI and maybe kind of to wrap up this conversation within news and publishing, what areas are you most excited about looking forward into how generative AI could be used within this kind of the sector within SEO?

Yeah, I'll be honest, I'm probably not the right person to ask that because I would call myself slightly conservative in my approach to SEO. I'm sort of always in a wait-and-see mode whenever a new technology comes around and a new trend happens. I try not to get too excited about anything new in the industry because maybe it's because I've been around for so long. I've seen all these, these hypes come and go so often. Everything is going to change SEO. We're going to do everything radically different now that this technology or that technology has been around. And It generally doesn't happen. SEO is constantly evolving. The web is constantly evolving. There hasn't been a major revolution in the web in, well, forever, if you ask me. The web has always been more of a constantly, gently evolving ecosystem with very few radical pivot moments. And I think generative AI was initially thought to be one of those radical pivot moments, but I see it more as an evolution as well. And I myself am quite curious to see where it's going to go. And I'm notoriously bad at making predictions, so I think I'll refrain from doing that and just say that, you know what, I'm just going to keep my eyes open. I'm going to try and keep an open mind and an open visor and see what this new technology can bring us. And I hope it's going to be mainly positive, but I'm also aware that technology in and of itself doesn't always have positive impacts. So I think we as a whole need to be aware of the fact that sometimes The technologies we build and unleash on the web or on the world in general may need to be safeguarded a bit. And I think we already are having interesting conversations about putting guardrails on AI. So I'm quite encouraged by that particular development.

Spoken like a true veteran. I also share the same belief. And I think we've seen it a lot with all the talk and kind of discussions and even a lot of the concerns around Google SGE. And for me, I think at least in the last 5, 10 years, it's like a lot of the fundamentals within SMO and the fundamental principles and concepts are still intact. We're just applying different tools in our approaches on execution, but it's not necessarily, at least from my perspective, fundamental shift in how search or how data retrieval systems operate in general, but it's more of Yeah, how you're adopting them into your tactics. But I mean, share the belief too that it is interesting and it is good to always be keeping an eye out, to be measuring this, to be willing to embrace the changes, but not necessarily overreacting to those. And with that, that wraps up this episode of the Voice of Search podcast. Thanks again to Barry Adams, specialized SEO consultant for news and publishers at Polemic Digital. In part 2 of this interview, which will be published tomorrow, Barry and I are going to continue the conversation and dig a bit deeper into SEO for news publishers. If you can't wait until the next episode and you would like to learn more about Barry, you can find a link to his LinkedIn profile in our show notes, or be sure to subscribe to his newsletter at SEO for Google News for regular updates from Barry on SEO for news publishers.

Okay. Thanks to Tyson Stockton, our guest host. If you'd like to get in touch with Tyson, you could find a link to his LinkedIn profile in our show notes. You can contact him on Twitter where his handle is TysonStockton. Or if your team is interested in SEO consulting or organizational education, you can always head to their company's website, which is previsible.io. That's P-R-E-V-I-S-I-B-L-E dot I-O. Just one more link in our show notes I'd like to tell you about. If you didn't have a chance to take notes while you were listening to this podcast, head over to voicesofsearch.com, where we have summaries of all of our episodes and contact information for our guests. You can also subscribe to our weekly newsletter, and you can even send us your topic suggestions or your marketing questions, which we'll answer live on the show. Our show. Of course, you can always reach out on social media. Our handle is Voices of Search on LinkedIn, Twitter, Instagram, Facebook, or you can contact me directly. My handle is Ben J. Shapp, B-E-N-J-S-H-A-P. And if you haven't subscribed yet and you want a daily stream of SEO and content marketing insights in your podcast feed, we're going to publish an episode every day during the work week. So hit that subscribe button in your podcast app and we'll be back in your feed tomorrow morning. All right, that's it for today. But until next time, Until next time, remember, the answers are always in the data.