AI or Human: Creating Helpful, Reliable, People First Content
Table of Contents (Click to show/hide)
Active User Rate
Return on Investment
Creating Quality Online Content in the Era of AI
A lot has happened in the era of AI-generated content lately. Most notably, Bing is now ChatGPT-integrated and Google announced its Bard AI search function. Understandably, this has raised concerns within the industry.
So, how can you and I keep creating ranking content in this heavily AI-focused era? Is it possible? The answer is yes, and I'm going to show you how.
The Concerns Around AI-Generated Content
AI's jaw-dropping capacity and generated content are hotly debated in the industry. People are worried that it could take over manual work, ridding creators of their jobs.
I consider them valid concerns — the AI results can be better than long-winded, human-written content. But after corroborating Nicole Farley's "test drive" of the ChatGPT-integrated Bing, other people's experiences with the bot, and my own play around, I can safely say there are still some major issues.
The rise of AI journalism, which can write news stories faster, cheaper, and without bias. The concern is not only that AI-generated stories will replace human journalists but also that AI can write news stories without needing to rely on human journalists to fact-check or cite sources, leading to a potential decline in the quality of journalism. Publishers must find ways to make their content more engaging and add an entertainment factor, incorporating personality, humour, and creativity into their content, something that machines cannot replicate at this point in time. Despite the rise of AI journalism, human journalists still have a crucial role to play, bringing insight, context, and empathy to news stories.
As Barry Schwartz from Search Engine Roundtable puts it,
"Maybe we need to add an entertainment factor. Maybe we need to add personality and humor, and entertainment into our content, above and beyond what a machine can do at this point in time."
So let's get creative, have fun with our content, and show our readers that we're more than just a bunch of algorithms spitting out facts.
Not only does it get information wrong and display biases, but some of the results could be damaging. In fact, several people have had alarming run-ins with the bot when using it for prolonged periods.
And from a purely technical side, there are some kinks to work out (like links that don't actually link to anything).
How to Ensure High-Quality Well-Ranking Content in The Era of AI
With those concerns in mind, Google has Set The Record Straight by offering additional guidelines and clarification surrounding its thoughts on all content, AI-generated included.
The company has always rewarded high-quality, helpful content. And that's here to stay as it continues focusing on articles' standards over production methods.
Utilizing automation like AI to create content with the sole goal of manipulating search engine rankings violates Google's policies. The company will keep using SpamBrain and other systems to fight such practices.
But, it still recognizes that not all AI use is spam — publishers automate useful information like weather forecasts, sports scores, and transcripts. More importantly, like what I just quoted from Barry Schwartz, AI consumes and output amazing and timely answers in its chat interface based on not just what publishers are posting on their own sites but also referencing posts on Twitter from experts in the field.
So, here's how you can hit Google's high content bar regardless of your content production methods:
Writing People-First Content
Whether Google is ranking AI-generated content or manually created pieces, its systems and rankers are looking for people-first helpfulness.
Writing for real readers is essential. Keeping this mindset for all content will inherently appeal to search engines, putting your website on the map. After all, Google's algorithms are rooted in user interactions.
If you're just starting out on the SEO journey, it can be difficult not to get sucked into overusing keywords. But I have a lot of great tips for writing reader-friendly content right here — and yes, interlinking related articles is one of them.
Familiarising Yourself with Google's E-E-A-T Guideline
I have an in-depth guide (case study included) on Google's E-E-A-T guideline that I encourage you to peruse for complete details. But here's a brief introduction for your convenience.
Google's automated systems utilize various factors to rank A-star content. Once they've identified relevant pieces, they'll use a mixture of methods to determine which articles demonstrate the most:
- authoritativeness; and
otherwise dubbed E-E-A-T. (Note: YouTube SEO also applies this visitor-first principle)
Self-assessing your content from an E-E-A-T perspective helps guide your edits to ensure criteria alignment, ultimately putting your site on the first page.
Asking "Who, How, and Why" for All Your Content
Evaluating your content with "who, how, and why" questions helps you stay on course and align with Google's ranking criteria. I use it as a quick-and-easy way to near-on guarantee adherence to the E-E-A-T guideline and ensure I'm creating people-first content, AI-generated or otherwise.
The Datasets Used for AI-Generated Content
Understanding the data source allows you to see that Google isn't entirely against automated content — it provides helpful answers for some topics. But Google's Bard AI is somewhat shrouded in secrecy.
Although, the company announced that it's based on the LaMDA language model comprised of the following mix:
- 12.5% C4-based data
- 12.5% English language Wikipedia
- 12.5% code documents from tutorials, Q;A sites, and more
- 6.25% non-English web documents
- 6.25% English web documents
- 50% dialogues from public forums
C4 was developed in 2020 by Google. The dataset is based on open-source Common Crawl data.
Common Crawl is a non-profit organization that scours the internet every month, making free datasets for anybody to use.
In Bing AI, the system leverages advanced algorithms and machine learning techniques to understand user intent, context, and preferences. It also uses natural language processing to interpret and analyze queries more accurately, helping to deliver more relevant and precise search results.
One of the key components of Bing AI is Prometheus, a deep learning platform that enables Bing to build, train, and deploy machine learning models at scale. Prometheus provides a flexible and efficient way to build and deploy models across a range of applications, including web search, image search, and voice search.
Bing Orchestrator is another key component of Bing AI, which helps to manage the flow of data between different parts of the system. Orchestrator ensures that data is passed between different components in a consistent and efficient way, allowing Bing to process large amounts of data quickly and accurately.
ChatGPT, a large language model trained by OpenAI, is also used in Bing AI to improve the quality of search results. ChatGPT helps to understand user queries and provides relevant search results based on the user's intent. ChatGPT is also used to generate natural language responses to user queries, making search results more conversational and user-friendly.
Overall, the integration of Prometheus, Bing Orchestrator, and ChatGPT enables Bing AI to provide more accurate and relevant search results, while also improving the user experience for people searching the web.
ALT text: Prometheus model architecture diagram showing the flow of data from input to output through various layers and modules
Title: Prometheus Model Architecture Diagram from Bing AI, Source: blogs.bing.com
Identify Intent - Content Opportunities from User Behaviours Using Search & Chat
Merging search and chat to create a seamless experience for users, applies not just search engine, but all businesses like ecommerce, B2B and more. Based on user behavior observations, Bing found that users tend to use chat when they are exploring and have a vague idea of what they want. This is what marketers call the "upper and mid-funnel." In contrast, users who know exactly what they want and have a specific query tend to use search to gather information.
For instance, users who are not sure about the best TV model that provides the highest energy rating at a good price may use a chat approach. This indicates an opportunity for content creators to focus on intent-based content creation. For upper to mid-funnel content, creators can use natural, descriptive, question-related language to help users understand their needs better. For lower funnel content, creators can use technical, research-based, analytical, and well-cited language to create more detailed and comparative content.
The Bottom Line: Quality is King
From everything I've researched and discussed today, there's one primary standout requirement for all content — quality. Google will continue upholding its reputation of rewarding helpful, people-first content, giving you clear guidelines to ensure your site gets noticed.