We have already heard that Google had Bard ready months before OpenAI introduced ChatGPT, but the company reportedly decided not to launch it. Thibault “Tibo” Sottiaux, a former Google AI researcher who is now with OpenAI, claims that Google internally developed a conversational AI nearly a year before ChatGPT made its public debut.
The discussion started on X (formerly Twitter), where AI engineer Cheng Lou referenced an old interview with Google Chief Scientist Jeff Dean, in which he said the company had an internal chatbot but believed it wasn’t more useful than traditional Google Search.
Responding to the post, Sottiaux claimed he worked on that project during his time at Google. As per him, the chatbot, internally known as LMChat before receiving another codename, was functionally similar to ChatGPT long before OpenAI launched its product.
Sottiaux further alleged that Google decided against releasing the chatbot because of concerns that such a product can undermine its core search business. He also claimed that DeepMind faced restrictions on launching products that might compete with Google’s existing services.
Also read: After OpenAI, Anthropic says Claude AI accidently hacked other companies: Here is what happened
Ironically, Google laid much of the foundation for today’s generative AI boom. In 2017, its researchers published the landmark paper Attention Is All You Need, introducing the Transformer architecture that powers modern large language models. Even the “T” in ChatGPT stands for Transformer.
Before ChatGPT’s launch, Google had already developed advanced AI models such as Meena, LaMDA and PaLM. However, most of the models remained internal or were released only in limited forms due to the concerns around privacy, misinformation and safety.
After ChatGPT’s explosive launch in late 2022, Google accelerated its AI efforts by introducing Bard, which was later renamed as Gemini. Today, both ChatGPT and Gemini remain among the world’s most widely used AI assistants, alongside Anthropic’s Claude.