AI for SEO in 2023: are the machines finally taking over?

Times are tough, and some are turning to AI tools to cut costs for their content creation and SEO. But will it pay off? What are the opportunities of AI for SEO – and what limitations and risks do you need to watch out for? Hear our expert insights and opinions, and how best to approach AI in 2023.

2023 was always going to be a tough time for many businesses. Battling economic recession and the ongoing impacts of a pandemic, many are looking to cut costs wherever possible – cutting back marketing budgets and even staff. On top of that, changes to privacy and data laws and regulations and the shift to Google Analytics 4 are making the digital landscape (and paid advertising) more complex. We’re seeing organic content prioritized as businesses seek to maintain their presence at a lower cost. So, it’s no surprise that, with the recent transformative advancements in the area, some companies are turning to artificial intelligence (AI) as a cost-saving choice for content creation.

But with a more organic and content-focused marketing approach, it's more important than ever for businesses to have a strong search engine optimization (SEO) strategy, so their content ranks well and will drive traffic to their website and get more exposure for their content. AI tools are being adopted there too – just ask any LinkedIn marketing guru, and they’ll tell you that, with the right prompts, OpenAI’s ChatGPT and its eerily advanced capabilities can even offer content that’s optimized for search.

So, has the moment finally come – are the machines taking over? Should you, as a marketing manager or CMO, be running out to subscribe to these tools immediately? And should we, as digital marketers, copywriters and SEO specialists, start looking to retrain in a new field? As you might suspect, the answer is not quite a simple as that.

Let’s talk about some of the current opportunities AI tools offer when it comes to SEO – their strengths and weaknesses, and some recommendations on the best ways to approach this new technology in order to get ahead.

AI-generated content and SEO

First, let’s talk about one of the internet’s current favorite topics – generated predictive text. Tools like ChatGPT use language-predictive algorithms to create content on a given topic. The advances in this area recently are truly mind-blowing – with quality content able to be produced on almost any given topic at the click of a button (or, in this case, the asking of a question).

These low-cost tools are true content machines when it comes to volume, producing copy on almost any topic to a fairly high language level. They even allow easy iteration to improve on aspects of the content and writing as needed, and with an intuitive and user-friendly chat interface, are easy for almost anyone to use. But there are a few downsides that must be taken into account before you jump on the GPT bandwagon.

Quality, SEO-optimized content still needs extensive human intervention – both at the input and output stages. You need to provide the best input and information for the content you want, and clear prompts to ensure the angle, content included, tone and style of the writing will be in line with your brand and/or media. You also need to iterate and edit afterward to check for quality, accuracy and relevance for your audience. Of course, all this work may negate much of the time and therefore cost advantage gained!

Using AI-generated copy also comes with a guaranteed lack of originality – the nature of predictive text means it’s using an algorithm to guess what word should come next, based on a ‘rough average’ of the information it’s been taught. For those who are curious, the New Yorker have an interesting take using a metaphor of Chat-GPT as a blurry image of the web – basically a ‘rephrasing’ of the internet – and all the mistakes and fabrications that come with that process. It’s far from flawless. Even OpenAI clearly state ChatGPT has limited knowledge and “may also occasionally produce harmful instructions or biased content”.

Being based on their machine learning materials, it’s also not possible for these tools to come up with any completely new concepts or expressions. So, if you’re happy to sound like everyone else, repeating what everyone else is saying about a topic, it’s great. But if you want to stand out – either with your branding, style, tone or angle, or as a thought leader – you’ll need to think outside the (chat) box.

AI-powered content: the good, the bad and the policy-violating

In my opinion, using AI GPT tools to create business content is a lot like using Wikipedia for a research article. It can be a great place to start, to get inspiration and ideas on what kind of thing you might talk about, or where to research – but not be relied upon as a trustworthy source on its own. It can, however, be an excellent tool for specific tasks.

They can be great at helping to optimize existing content for specific, relevant SEO keywords (researched separately using SEO keyword tools and provided as part of the prompts). It doesn’t replace the need to do your keyword research for volumes and terms relevant to your target audience though! It can also be very effective at helping to summarize your content for effective metadata descriptions, or giving some inspiration and ideas for heading options that include your keywords.

For one of our clients, for example, we were able to successfully create short text descriptions for use with internal links on their pages. With some good prompts and using the content from the pages to guide the input, we were able to craft useful, optimized texts. But when it came to another client in the finance industry, it was a lot more challenging. In areas like finance, to establish trust and expertise, it’s important to take a responsible approach and tone when it comes to talking about money. After all, not everyone can or should be encouraged to take a loan.

But when we tried to use an AI tool to write meta descriptions for a financial client’s web pages, all we received was text for comparison sites (“Compare and find the best loan company here”) or for low-cost or guaranteed loans (“Get the best loan with low monthly payments today”). These ‘sales-pitch’ suggestions didn’t align with the search intent we were optimizing for, nor with the tone we wanted to communicate. It was taking more time to try to improve and rework the prompts than it would have taken to just write the text from scratch.

Attribution, transparency and trust

There are other concerns too when it comes to using AI for content production. According to OpenAI’s content policy and terms of use, users (even non-paying ones) do own and may use, even commercially, the output they create. They do, however, “encourage you to proactively disclose AI involvement in your work” and expressly state you “may not mislead people…[or] tell people that the work was entirely human generated”. This can pose an issue for thought leadership content or expert advice and information – a financial advice blog, for example. Best practice for this type of content would be to attribute the writing to an actual person, to build trust and authority. And in industries such as finance, healthcare or governmental policy and communications, the human element is crucial.

In a well-documented case, formerly-venerated news site CNET used AI to generate over 70 finance stories packed with lucrative affiliate links – and in some cases, harmful errors – for disastrous results. Not only did this content (not clearly disclosed as AI-generated or in some cases attributed to human authors in the bylines) confuse and misinform readers while profiting off them, it caused great concern for the employees around the journalistic and editorial integrity of the company and its executives. The revelation of the scandal damaged not only the brand itself, but customer trust in the industry.

Then there’s the Google algorithm to contend with. With its own AI search on the way, the search powerhouse is well-placed to analyze and identify AI-generated or optimized content, which may not align with their quality rater guidelines and may even violate Google’s spam policy if the content is not people-focused, valuable and transparent. This could end up in an opposite result to that intended - with your website being penalized by Google and losing valuable traffic and rankings.

What does the future hold for AI and SEO?

From the success stories of AI in general across industries worldwide, we know there are two particular areas in which AI truly shines – the analysis of large quantities of data to identify patterns and predict trends, and the automation of repetitive tasks, freeing up employees to spend their valuable time on more important, creative or people-focused activities. This is where I think we can expect to see AI tools really soar when it comes to SEO – in keyword trends and predictions. Bing and Google are already launching AI-powered search, so it’s likely the insights learned from this new technology will also shape the future of search rankings.

It also raises some burning questions that it will be interesting to see the answers to:

Will we soon need to optimize our content for AI algorithms?

Instead of optimizing for Google’s search ranking algorithms, will we be optimizing content for selection and featuring in AI search? Will they be the same, or different? While Google’s current rater guidelines maintain the people-focus prioritization, it will be interesting to see how this plays out in reality once AI search is implemented.

Will sources be incorporated in content generation tools like ChatGPT?

This would be a useful addition, not only to enable transparency in terms of the accuracy of information and the critical evaluation of sources, but also to add an element of explainability when it comes to where the algorithms are pulling the information from for certain queries.

Will paid content or sponsorships be introduced to AI-generated content?

Will companies be able to pay to have their content featured more in the text or results generated about a topic? This is particularly interesting if we start to see sources incorporated as in the above.

Will we finally see voice search and assistants come into their own with AI?

With extensive information and more natural, intuitive communication (becoming ever-smarter over time as the machine learning algorithms learn from our behaviors), maybe voice services like Google Assistant, Siri and Alexa will finally become to search what the early hype said they would be – a force to be reckoned with, and a useful day-to-day tool (beyond turning on and off your lights).

Will Bing gain more market share as a result of their AI search launch?

If so, SEO and marketing professionals will have to start looking into how our companies or clients are performing there too. It might be time to start looking into Bing webmaster tools!

What will the future of SEO look like with AI?

Will we have our content AI-managed and automated to adapt to user search trends, optimizing for keyword trends and search volume over time? Will the best AI-optimized content win?

We’re also seeing some use of AI tools in the production of code for technical optimization in SEO, as Aleyda Solis mentions in her blog (formulas, rules, tags and markups for example), and also in the translation of content to other languages. While we would say the quality of the output for these two areas are still not reliable or consistent enough to be used professionally without expert review, it is an exciting area with great potential for future use.

Whatever the future may hold, it’s a fascinating time of transition for SEO, content and digital marketing – we’re really witnessing a transformation that will shape the years to come.

So, has AI replaced us when it comes to SEO?

Generally, I’d say no – the human element of creativity and originality is still too important when it comes to quality content, personalization and understanding (and speaking to) your audience. Leading companies and CMOs should see this as a greater opportunity to stand out, rather than blend in to the beige.

But I will add – things will be changing a lot, and marketers who ignore AI technology and the tools and opportunities it brings will absolutely be left behind. Don’t be afraid – be curious. Read up on the latest tech, learn about the new tools, and get a better understanding of how AI should (or should not) be used to meet your or your client’s needs.

Otherwise, perhaps it’s time to start updating your LinkedIn profile. Maybe ChatGPT can help?

By Fanny Pettersson SEO Consultant