OpenAI's New AI Model Puzzles Experts with Multilingual Reasoning

OpenAI's newest AI model is switching languages to Chinese and others while reasoning, puzzling users and experts

Serving tech enthusiasts for over 25 years. TechSpot means tech analysis and advice you can trust. WTF?! OpenAI's latest AI model, o1, has been displaying unexpected behavior that has captured the attention of both users and experts. Designed for reasoning tasks, the model has been observed switching languages mid-thought, even when the initial query is presented in English. Users across various platforms have reported instances where OpenAI's o1 model begins its reasoning process in English but unexpectedly shifts to Chinese, Persian, or other languages before delivering the final answer in English. This behavior has been observed in a range of scenarios, from simple counting tasks to complex problem-solving exercises. One Reddit user commented, "It randomly started thinking in Chinese halfway through," while another user on X questioned, "Why did it randomly start thinking in Chinese? No part of the conversation (5+ messages) was in Chinese." The AI community has been buzzing with theories to explain this unusual behavior. While OpenAI has yet to issue an official statement, experts have put forward several hypotheses. Some, including Hugging Face CEO Clément Delangue, speculate that the phenomenon could be linked to the training data used for o1. Ted Xiao, a researcher at Google DeepMind, suggested that reliance on third-party Chinese data labeling services for expert-level reasoning data might be a contributing factor. "For expert labor availability and cost reasons, many of these data providers are based in China," said Xiao. This theory posits that the Chinese linguistic influence on reasoning could be a result of the labeling process used during the model's training. Another school of thought suggests that o1 might be selecting languages it deems most efficient for solving specific problems. Matthew Guzdial, an AI researcher and assistant professor at the University of Alberta, offered a different perspective in an interview with TechCrunch: "The model doesn't know what language is, or that languages are different. It's all just text to it," he explained. This view implies that the model's language switches may stem from its internal processing mechanics rather than a conscious or deliberate choice based on linguistic understanding. Tiezhen Wang, a software engineer at Hugging Face, suggests that the language inconsistencies could stem from associations the model formed during training. "I prefer doing math in Chinese because each digit is just one syllable, which makes calculations crisp and efficient. But when it comes to topics like unconscious bias, I automatically switch to English, mainly because that's where I first learned and absorbed those ideas," Wang explained. While these theories offer intriguing insights into the possible causes of o1's behavior, Luca Soldaini, a research scientist at the Allen Institute for AI, emphasizes the importance of transparency in AI development. "This type of observation on a deployed AI system is impossible to back up due to how opaque these models are. It's one of the many cases for why transparency in how AI systems are built is fundamental," Soldaini said.

Why Does ChatGPT's Algorithm 'Think' in Chinese?

OpenAI's new reasoning model is doing weird, unpredictable stuff. OpenAI recently unveiled a new algorithm, o1, which the company calls its "reasoning" model. The idea behind the algorithm is that it spends "more time thinking" before it responds, thus delivering better answers. However, the algorithm appears to "think" not just in English but in several different languages. Web users recently noticed that the program was randomly displaying Chinese characters, as well as code that appeared to be written in several other languages. Most people will only pay attention to the final output that ChatGPT produces, but users have the option to look at how the model is reasoning its way to an answer. That's where many people noticed that the LLM had begun to incorporate Mandarin and Cantonese into its process. "Why did o1 pro randomly start thinking in Chinese?" asked Rishab Jain, via X. "No part of the conversation (5+ messages) was in Chinese... very interesting... training data influence." Prior to that, on January 5th, another web user, Nero, wrote on X: "uhmm, why is my gpt o1 thinking in chinese, lol." They tagged both OpenAI and ChatGPT in their message but received no reply. The obvious answer would appear to be that the algorithm was likely trained on large tranches of Chinese data, and, thus, that data is influencing the algorithm's output. "Most probable explanations - certain languages might offer tokenization efficiencies or easier mappings for specific problem types," said Rohan Paul, an AI engineer. "So, o1 may be switching languages because its internal representation of knowledge finds that using Chinese can lead to more optimized computation paths when handling certain problems." Another online commentator, Raj Mehta, gave a similar explanation: "o1, like many large language models (LLMs), operates in a shared latent space where concepts are abstract, not tied to specific languages. It might "reason" in the language that maps most efficiently to the problem at hand." Gizmodo reached out to OpenAI for comment but didn't receive an immediate explanation. There's plenty of speculation flying around about why this is happening, though TechCrunch interviewed Luca Soldaini, a research scientist at the Allen Institute for AI, who seemed to offer the best response to the conundrum. Soldeini said that, due to the opaque nature of the company's algorithms, there's really no way to know why the program is behaving in the way that it is. “This type of observation on a deployed AI system is impossible to back up due to how opaque these models are,†Soldaini told the outlet. “It’s one of the many cases for why transparency in how AI systems are built is fundamental.†Indeed, the black-box nature of corporate AI algorithms is particularly ironic given OpenAI's self-stated "open" mission of transparent technological development. "OpenAI isn't so open, which means that when its algorithms do weird things like spout Chinese, all we can really do is scratch our heads and wonder why."