Any readers of this article will undoubtedly remember the state of the internet in 2014: a sea of low-quality, mass-produced, ‘clickbait’ content produced by content farms for the sole purpose of grabbing attention, traffic, and social shares (and of course ad revenue).
Arguably, this type of content never went away, although we did become a tad more wise to it, and social networks and search engines took measures to reduce its visibility: Facebook adjusted the NewsFeed algorithm to prioritise links that featured additional context, and downgrade those that were buried amid status or photo updates, as well as those that users ‘bounced’ away from quickly; and Google’s Panda updates targeted poor-quality content that gamed the search algorithm.
A decade ago, content farms and clickbait sites at least needed to pay writers to churn out content, even if it was low-effort content. In 2023, however, generative AI tools make it possible to mass-produce text and imagery almost at the click of a button. This can be employed to a beneficial end, but there have long been fears about these tools enabling a wave of AI-generated spam, and all the signs indicate that it’s already here. Recently, a tweet by user @generalslug went viral showing an article with hilariously incoherent AI imagery that ranked for queries related to hair waxing:
wanted to know the ideal hair length for waxing and was immediately shown a 100% AI-generated article, including images pic.twitter.com/ABoBJhpXnV
— SLUG (@generalslug) June 14, 2023
In May, NewsGuard, a journalism tool that rates the credibility of news and information websites and tracks online misinformation, warned that “A new generation of content farms is on the way” in a report that identified 49 different websites across seven languages that “appear to be entirely or mostly generated by artificial intelligence language models designed to mimic human communication”, and “often fail to disclose ownership or control”.
Google’s August 2022 Helpful Content Update was designed to mitigate the issues presented by spammy content, including – but not limited to – automated content, and Google is reportedly debating where AI content fits into its E-E-A-T search quality guidelines. But will this be enough? I turned to three search and AI experts to get their thoughts on the new AI content wave and what it might mean for users’ experience of, and trust in, search.
Google won’t let AI spam compromise its business model
Beth Nunnington, VP Organic Media at Journey Further:
“Google isn’t going to risk its business model, which relies on the quality of the SERPs. It will either quickly get better at understanding AI generated content – or you’ll likely stop seeing this type of results.
“Google may have been forced to launch this technology earlier than they had planned due to ChatGPT and other rivals starting to implement AI-generated SERPs – but with the resources Google has available to them and the years of work developing AI behind the scenes – we expect this to get better, fast.
“If AI content regularly appears in results that is seemingly false or misleading, users are unlikely to engage. That’s why I believe that this won’t continue to happen, and Google will be working hard behind the scenes to ensure its algorithm still only publishes the most relevant and useful content.”
Users will flock elsewhere for trustworthy information
John Clarvis, Data & Insights Director at The Kite Factory:
“I predict that there is a lot more of this to come. Spam is and always has been a numbers game, and generative text and images speed up the process for spammers even more.
“As it stands, the tech to create spam has outpaced the tech to detect it, and I expect that search engines will be rapidly developing ways of downranking and excluding low quality AI content. Whichever one can provide the most accurate and useful information will then dominate the market. Google has stood out over time as its search results were head and shoulders above the rest, however AI spam has opened up an opportunity for Bing et al to outperform Google and steal back some share.
“Search is an everyday part of people’s lives … and users have become accustomed to finding their desired results quickly and efficiently. If all they get in return is AI generated clickbait or spam, we may see consumers flock to other search engines or personally trusted platforms like social media, Reddit or Mumsnet for their information.
“But this result wouldn’t just stem from AI-created content, but the lazy application of AI to create low-quality, click-bait spam. Humans can also spam and obfuscate information to play the SEO game – searching for recipes is a prime example of humans padding out content to rank better. Whether human or AI, the end result is the same – annoying, irrelevant information that ultimately discourages people from clicking.”
Human supervision of AI outputs is vital
Lemuel Park, co-founder and CTO at BrightEdge:
“I expect to see a lot more of this, especially on social media and sharing sites, especially from standalone tools and applications. This marketplace is now flooded with nearly 500 new entrants offering all types of output and use cases. This expands beyond content generation and can include personal and recreational users, gaming, music, coding and development tasks, audio, imagery, video, and podcasts.
“Humans are experimental in nature, so new tools, especially those that give visual or engaging outputs, allow people to experiment and share their experiences regardless of whether it is right or wrong.
“However, regardless of whether the outputs are from a standalone application or a search engine, Generative AI also has limitations. … With countless individuals experimenting with Generative AI, caution is also essential. All language models can produce meaningless outputs. Ensuring human supervision is at the front and end of any process where you want to publish content is vital.”
An opportunity for brands to stand out and build trust
“AI-generated content will lead to an increase in generic, unimaginative content, but this creates a huge opportunity for brands to stand out by making original content that is rooted in brand truth and insight.
“Brands that follow this approach and create useful and relevant content that their audience wants to read, which aligns with their brand guidelines and messaging, will succeed. This will help them successfully engage with their target market and meet the expectations of search engines (Google).”
“Brands will be more important than ever, and the potential future of senseless AI generated content is a great opportunity for brands to build credibility. If a brand stands out in search for its quality of content then people will mentally uprank it in their decision to click. That being said, using AI to save time and generate content isn’t necessarily a bad thing for brands. In fact, it’s really fun and can throw out some brilliant insights and novel solutions to problems.
“Everything a brand does online reflects on the brand, so ensuring quality is paramount. While distinctly analogue, proofreading is still crucial for anything that a brand puts out.”
An opportunity for SEO teams to shine
“Generative AI presents an excellent opportunity for SEO teams to shine and optimize content for better customer experiences.
“For content to perform well, it must have a clear purpose. It is essential to differentiate your content from the rest to increase discoverability and online visibility. Re-evaluate the reasons for creating content, the methods used to make it, and the intended audience. Content must align with your target audience’s intentions and solve their needs and desires.
“As Google looks to display a variety of content from various sources and users and refines its content system to offer more valuable and relevant information – follow its Helpful Content Update (HCU) guidelines, especially as these are being refined to recognize personal and expert perspectives.
“Collaborate closely with your SEO team to collate user and conversational intent information – a key trigger for Generative AI results. Utilize data-led insights to understand how users interact with search engines and websites. SEO insights will guide you on what content is performing well and what types and formats work best – for you and your competitors.
“And exercise patience – generative AI technology is still in its infancy, and it’s important for content professionals to approach it cautiously. Ensure any auto-generated content has human supervision and an editor.”
Does it matter how it was authored – as long as the quality is there?
“Ultimately, people care that the information is high quality and useful. It doesn’t matter if it’s been made by a human or an AI.
“It’s the same as manufacturing: nobody cares that their car has been built by robots rather than by hand, so long as the car works.”