ChatGPT Language Switching Issue

The first report surfaced on March 26th, when a user noticed ChatGPT had spontaneously dropped a Russian word into an otherwise normal conversation — a word they believed translated to something like "alien." Odd enough to mention, not odd enough to panic about. Then, three days later on March 29th, a second user reported their entire chat session had shifted into what appeared to be Russian, with no instruction given and no obvious trigger. Two data points, similar flavour, a few days apart. That is either a coincidence or the beginning of a pattern — and right now, we genuinely cannot tell which.

If confirmed, here is what this means. A language model inserting unsolicited foreign language output is not a trivial glitch — it suggests something is misfiring in the model's language detection or response generation layer, possibly tied to training data bleed or a misconfigured system prompt at the infrastructure level. For users who rely on ChatGPT in professional or multilingual environments, unpredictable language switching would erode trust in outputs fast. The second-order concern is broader: if the model is making autonomous decisions about output language without user instruction, it raises quiet questions about what else it might be deciding without being asked.

Watch for a spike in similar reports across Reddit, X, or OpenAI's own community forums — and pay particular attention to whether affected users share any common settings, browser configurations, or account regions that might point toward a reproducible cause.

NewsHive monitors these sources continuously. All signal titles above link to the original reporting.

Intelligence by NewsHive. Need help navigating what this means for your business? Contact GeekyBee →