Google's Pre-ChatGPT AI: Why It Wasn't Released

Google reportedly built a ChatGPT-style chatbot before OpenAI, but chose not to release it publicly. Computer scientist Jeff Dean stated that Google's leadership held back the chatbot due to concerns about accuracy and safety. He mentioned the model sometimes provided incorrect information from a search perspective and had safety issues, including occasionally saying offensive things. Dean noted that Google employees used this internal chatbot during the pandemic, finding it a "nice partner" for conversations. The company also didn't fully realize its potential for tasks beyond traditional search queries, such as writing notes or summarizing text. This highlights the early challenges and considerations in developing conversational AI.