Is there an official OpenAI policy for inconsistent Pulse language output in a specific non-English language?

OpenAI_Support:

I’m not aware of a public policy or formal statement specifically covering inconsistent Pulse language behavior.

Language output can be influenced by a mix of factors like device settings, app language, and context, so occasional inconsistencies can happen, especially in non-English scenarios. I haven’t seen this called out as a widely acknowledged issue specific to Pulse.

If this can be consistently reproduced across users, it would typically be treated as a system-side behavior and looked into further. Sharing a few concrete examples would help clarify whether this is a broader pattern.

Thank you for your reply.

However, I must respectfully disagree with the explanation that “language output can be influenced by a mix of factors like device settings, app language, and context.”

My situation is as follows:

• My ChatGPT account language is explicitly set to Korean in the settings.

• Both my PC web version and my Galaxy S24 Ultra (Android app — always kept at the latest version) have Korean set as the system language.

• I submit queries in Korean approximately 99% of the time. Even on the rare occasions when I upload an English document for reference, I still write the actual question in Korean.

Despite all of the above, the daily Pulse notification is delivered in English on completely random dates with no correlation to content, context, or my query language. For example, in the period I have tracked (since March 21), it arrived in English only on: March 24, April 3, April 11, April 13, April 14, and April 15. All other days it correctly arrived in Korean.

This behavior completely ignores the account language I have explicitly set and appears to be decided by some internal, opaque logic that you yourself described as “a mix of factors.” By acknowledging that the output is not strictly following the user’s account language setting , your response essentially confirms that there is a flaw in OpenAI’s language-determination logic for Pulse.

This is not a user-side issue (device, app version, cache, or settings) — it is a clear system-side bug. As a ChatGPT Pro subscriber paying $220 per month, receiving a daily feature in the wrong language at random is extremely frustrating and undermines the value of the product.

I kindly ask that this be escalated to the engineering team as a reproducible bug rather than treated as occasional “inconsistency.”

Thank you for your attention. I look forward to a proper resolution.