AI-generated content needs to be fact-checked
Whether we like it or not, artificial intelligence is increasingly penetrating the lives of every Internet user. More and more often, its capabilities are used by specialists from different fields. But at the same time, the question of how high-quality the content created with the help of AI is, and whether it is possible to use the generated materials in one’s professional activities, is acute. Google specialists expressed their opinion on this matter in a recent episode of the Google Search Off The Record podcast. We would like to share with you the main conclusions.
The essence of the experiment by Google specialists
Members of the Google team that researches the intricacies of search algorithms conducted an experiment in April 2024. They took turns using Gemini to write texts for social media posts about technical SEO concepts. They were interested in two questions:
- whether factual errors will appear in the generated texts;
- Is human control necessary before publishing such posts?
Read on to find out what conclusions they came to.
Generative AI can make factual errors
The results of the experiment raised concerns among Google team members about the presence of factual inaccuracies in the generated materials.
Analyzing the content created by custom mobile app Gemini, experts named the main limitation of using such texts: “The biggest problem when working with AI-generated material is the use of facts that need to be carefully checked. This scares us, because we risk publishing material that is not true.”
Outdated SEO Tips Found
Google has raised concerns about an AI-generated tweet that suggested using rel=”prev/next” for pagination, a method that algorithms no longer support.
Gemini suggested posting the following tweet:
“Is pagination causing duplicate used to keep food and beverages content headaches? Use rel=prev, rel=next to follow Google’s algorithms. #technicalSEO, #GoogleSearch.”
SEO experts stressed that this approach has been used before, but it should be remembered that algorithms at the current stage will ignore this command. Errors in the generated text most likely appeared because outdated information was used in the AI training process.
before being published
During the discussion, the Google Searckind of usa lists guaranteeh Relations team noted that AI-generated content has great potential. But it is also necessary for people to fact-check the texts. This is about the responsible use of AI. Human oversight is necessary to prevent the spread of misinformation.
As the use of generative AI grows, remember that its results cannot be blindly trusted without verification by experts in the field.
How SEOs and Content Marketers Can Ensure the Accuracy of Generated Content
To ensure the accuracy of their publications, companies should:
- carry out a thorough check with the help of experts;
- clarify whether the content complies with current recommendations and modern practices;
- verify any AI data or recommendations with trusted sources;
- Stay up-to-date with the latest developments to identify outdated information that is used by AI.
Conclusions
AI-powered tools have the potential to help create and analyze content, but as Google’s experiment shows, a healthy degree of skepticism about using such services is warranted.
Blindly using generative AI can result in publishing outdated or harmful content, which can negatively impact your SEO performance and overall company reputation.
Search engines prefer high-quality and accurate information. Therefore, publishing unverified AI-generated content can worsen your site’s ranking in search engines.
It is vital to thoroughly fact-check AI-generated content and content with experts to ensure it is in line with modern best practices.