We saw recently that Reddit, the social forum platform with millions of posts and threads, started auto-translating its entire website using AI into multiple languages. 

This was not as an experiment, not on a few pages but at a scale of a million pages.

And during an earnings call, Reddit’s executive team said something that turned heads across the SEO community: “It’s totally sanctioned by Google.”

That raised eyebrows in the SEO world. 

Toggle

Free SEO Audit: Uncover Hidden SEO Opportunities Before Your Competitors Do

Gain early access to a tailored SEO audit that reveals untapped SEO opportunities and gaps in your website.

Analyze My Site

Because for years, Google’s take on automated translation was unequivocal: Don’t do it! It was considered “auto-generated content,” which generally fell under spam in Google’s book. 

So what happened? Why now? And what does this mean? 

How Reddit’s AI Translation Rollout Triggers a Policy Shake-Up in Google?

Why did Reddit do this? Simple: to reach more users around the world and get more traffic from non-English search results through Google which is the world’s biggest search engine.

So how was Reddit getting a pass?

Glenn Gabe, the well-known SEO consultant may have the answer to that. 

OK, I have an update on the Reddit AI translation situation. Google sent me a statement after I asked if they sanctioned AI translations for Reddit, if AI translations were now ok to publish, etc. Like I covered in my post about the subject, Google’s scaled content abuse spam… https://t.co/rwSdpRWFcL

— Glenn Gabe (@glenngabe) June 11, 2025

Upon reviewing Google’s official Search documentation for multilingual and multi-regional websites, Gabe noticed something subtle but major: Google also removed the section in their help docs on using the robots.txt to block all automatically translated pages.

There was no changelog, no blog post and no formal announcement from Google.

Gabe stated: “I had a feeling they would do that,” Gabe tweeted. “I think it was just overlooked initially, and now they’re updating based on broader AI use.”

It did not take long for the SEO world to start piecing the puzzle together and to draw the conclusion that Google was quietly shifting its stance on AI-translated content.

Shortly after, a Google spokesperson issued a clarifying statement that backed this up:

“Our policies do not strictly define content that has been translated by AI as spam. Our scaled content abuse policy mentions automated transformations, including translations, as part of the overall warning against creating large amounts of unoriginal content that provides little to no value to users.”

This is significant. Google is no longer saying “AI translation = spam.” Instead, the focus has shifted toward value to users and quality of content.

In other words, how the content serves the user now matters more than how it was created.

Why Does Google’s New AI Translation Policy Matters?

Let’s pause and think that AI is currently powering a content explosion. From e-commerce sites to global publishers, everyone is racing to scale up multilingual reach. 

According to a 2024 report by CSA Research, over 76% of online users prefer to purchase from sites in their native language. That is not just a preference but it is a conversion strategy.

Google Translate alone processes over 100 billion words per day and supports 249 languages. Add in tools like DeepL, ChatGPT and Meta’s NLLB models and the barrier to translating at scale is almost gone.

But the catch? Quality. Relevance. Context.

This is where Google’s updated stance becomes nuanced. It doesn’t say “all AI translation is okay.” It says AI translation that provides real user value is acceptable.

That distinction is critical.

From Blanket Bans to Balanced Rules Google’s Evolving Content Policy

Google’s Scaled Content Abuse Policy now refers broadly to: “Content that is generated at scale to manipulate search rankings and that doesn’t provide value to users.”

This includes AI-translated content if it is:

• Poorly localized
• Lacking cultural adaptation
• Spammy or misleading
• Simply duplicating existing content without nuance

However, if your translated content is:

• Helpful
• Context-aware
• Culturally relevant
• Properly indexed with hreflang and canonical tags

Then you are in the clear to rank in Google. 

This policy shift reflects a more mature and realistic view of AI in content production. It acknowledges that AI is a tool just like spell checkers, CMSs or design software, not a signal of spam on its own.Want the full backstory? At Stan Ventures, we have broken down Reddit’s AI translation rollout and Glenn Gabe’s findings in detail in this news. Reddit’s AI Translations Dominating Google: Why Is Google Silent, Asks Gabe. – Stan Ventures

What Can Go Wrong When Doing AI Translation?

Let’s not get too excited. Risks still abound:

• Literal AI translation fails: Idioms, humor, tone and emotion are often lost in translation. For example, translating “It’s raining cats and dogs” into literal Spanish or Hindi results in nonsense.
• Legal issues: Poor translations in medical, legal or financial sectors can lead to non-compliance or even lawsuits.
• Duplicate content traps: Poorly handled hreflang tags or publishing English and French pages with nearly identical AI output can confuse Google’s crawler and tank both pages.

Here is one example: Alibaba’s attempt to localize product listings for Amazon using basic machine translation resulted in product descriptions that were incoherent or unintentionally offensive and leading to flagged listings and negative ratings.

What are the Do’s and Don’ts for AI Translated Content

If you are managing a content-heavy website especially in industries like SaaS, media or e-commerce then properly translating your site is key to scaling globally without hurting your SEO. 

Here is what to do and avoid when using AI for multilingual content:

Do:

Implement Proper Multilingual SEO

Use hreflang tags, canonical URLs and proper site structure to help Google identify and index the translated pages for the correct regions and languages.

Post-Edit for Tone and Accuracy

Always review and refine the AI-generated translations with human editors or native speakers. This ensures the tone, intent and cultural nuances are preserved correctly.

Disclose AI Usage Transparently

Let users know if content was machine-translated and reviewed (e.g., a note saying: “This page was automatically translated and reviewed for clarity”). It builds trust and sets expectations. 

Use Trusted AI Translation Tools

Use high-quality tools like DeepL, Google Cloud Translation or GPT-based APIs. These offer better language models and accuracy compared to free, basic tools.

Don’ts 

Don’t Auto-Publish Without Review

Avoid publishing thousands of AI-translated pages without human oversight. This often results in poor grammar, wrong context or misleading information which Google may flag.

Don’t Confuse Translation with Localization

Translation is converting text to another language. Localization adapts it to a region’s culture, tone and expectations. AI can translate but localization still needs human touch.

Don’t Translate Without Cultural Sensitivity

Literal translations often miss local expressions or norms. For example, jokes or idioms may not translate well and could confuse or even offend users.

Don’t Ignore Hreflang or SEO Tags

Incorrect or missing hreflang tags can confuse search engines, causing duplicate content issues or ranking problems across international versions of your site.

So yes, Google is changing its translated content guidelines. But not to open the floodgates.

Instead, it is acknowledging that AI is now a critical part of the modern content stack, and rather than banning it outright, it wants to ensure it’s used wisely. This is about quality at scale, not quantity without value.