How to deal with Google’s new AI rules for Merchant Center

how to deal with google
Google has announced a significant update to its Merchant Center product data specifications regarding AI-generated content. As of April 2024, Google will  require merchants to disclose if their  product titles and descriptions contain AI-generated copy. In this post we look at myths and facts about genAI usage in product content.

Google has a new AI content disclosure requirement for sellers

Google has announced a significant update to its Merchant Center product data specifications regarding AI-generated content. As of April 2024, Google will  require merchants to disclose if their  product titles and descriptions contain AI-generated copy. 

According to Google, the move is intended to “enhance transparency and maintain the integrity of product information across Google’s shopping platforms.” Hmm.

But y tho?

Google could use this new information in three ways. 

First, it could improve its algorithm’s ability to distinguish between human and AI, enhancing the relevance and quality of search results. 

Two, it may influence consumer trust, as shoppers might prefer or avoid products based on the nature of the content’s creation. 

And C — while entirely speculative — Google could also use this data to develop new features or ranking signals specifically to distinguish between the nuances of AI-generated versus human-written content.

To disclose or not to disclose

Google itself can’t definitively tell if your titles and descriptions are AI-generated just by looking at the text. They may, however, use a combination of factors to raise flags and encourage you to use the new structured attributes:

  • Statistical Patterns: AI-generated text often follows certain statistical patterns. Google’s algorithms might analyze things like word choice, sentence structure, and the overall flow of the text to identify characteristics commonly associated with AI output.
  • Inconsistent Style: If your product data has a mix of titles and descriptions that seem very different in style or tone, it could raise suspicion, especially if some are particularly generic or keyword-stuffed.
  • Manual Reviews: Google might use human reviewers to analyze flagged content, especially for complex products or suspicious patterns.

Optimizing titles and descriptions through concatenation, keywords, and general content enhancement wouldn’t necessarily be considered AI-generated according to Google’s current Merchant Center guidelines. Here’s why:

  • Manual Intervention: You’re actively involved in the content creation process, making decisions and controlling the final output. This differs from AI, which operates more independently.
  • Rule-Based Techniques: Concatenation and keyword insertion are rule-based techniques, not involving the complex algorithms of AI that learn and adapt.

However, there are some nuances to consider:

  • Highly Formulaic Content: If your optimization process creates very generic or repetitive titles and descriptions across your products, it might raise a flag for Google.
  • External AI Tools: Some data feed tools might incorporate basic AI for content optimization. Double-check with your tool provider if they use any AI and how it functions.

Here are some tips on how to modify your titles and descriptions (with AI) to avoid them being flagged as AI-generated:

Prompt with a human touch:

  • Readability and Clarity: Prioritize user understanding. Write clear, concise, and informative titles and descriptions that accurately reflect your product’s features and benefits.
  • Unique Voice: Inject some personality! Use language that reflects your brand voice and target audience. Avoid overly generic phrasing or excessive keyword stuffing.
  • Specificity: Go beyond basic product details. Highlight unique selling points, user experiences, or interesting details that wouldn’t be readily generated by an AI.
  • Calls to Action: Encourage user engagement. Craft descriptions that entice users to learn more or consider purchasing your product.

Structure and Creativity:

  • Vary Sentence Structure: Mix up short and long sentences, avoiding a monotonous flow.
  • Bullet Points and Lists: Break down information for easy reading and highlight key features.
  • Questions and Answers: Frame some descriptions as questions users might have, then answer them directly.

Examples:

  • AI-Generated (aka generic): “Comfortable and stylish t-shirt made from high-quality materials.”
  • Human-Touch: This ultra-soft t-shirt drapes beautifully and keeps you cool all day. Available in a range of vibrant colors, it’s the perfect addition to your casual wardrobe. Pair it with jeans for a relaxed look or layer it under a blazer for a touch of effortless style.

The choice is yours

The goal is not just to avoid being flagged by Google but to create compelling, informative content that resonates with your customers. 

Unless written by its own AI and indexed somehow, there’s no definitive way for Google to determine if AI was used, but these tips will help your titles and descriptions appear more natural and user-focused. Use this opportunity to review your product feeds for high-quality, informative content that “speaks” to your target audience. 

Whether you use AI or an optimization tool like GoDataFeed or you write yourself, whether you choose to disclose or not, the choice is yours.

Share the Post:

Related Posts

How to Train AI for Your Brand

How to Train AI to Reference, Cite, and Crawl Your Brand

In AI-driven search, your brand’s visibility hinges on more than just traditional SEO tactics. You need to be a recognizable entity that machines can reference. If your brand lacks structured data, consistent naming, or a presence in trusted sources, you risk being overlooked or misrepresented. This is where reference engineering comes into play—transforming how your brand is perceived by AI. Discover how to build a robust identity that not only gets cited but also stands out in the crowded digital space. Don’t let your brand fade into the background; learn how to make it referenceable.

Read More
New SEO Stack

The New SEO Stack: Schema, Entities, and Answerability

SEO isn’t SEO anymore. It’s something else and the shift from keywords to entities is transforming how we approach content altogether. It’s no longer about matching phrases; it’s about clarity and structure that resonate with AI systems. Schema and answerability are redefining visibility and the meaning of “search results.” Your product pages must function as data objects, and internal linking is crucial for contextual signals. As we transition to Answer Engine Optimization (AEO), understanding how to make your content machine-readable is essential. Take a brief dive into the new rules of engagement and learn how to ensure your content is not just seen, but sourced.

Read More