Noam Shazeer was a former engineer on Google Brain, who was a key factor in the development of its language-AI technology. He is now the CEO of Character.ai, a company that has amassed nearly $200 million in funds for user-conversation with “characters” who possess multiple personalities. This revolutionary chatbot created renewed interest in generative AI. It sparked Microsoft to invest billions in OpenAI and consequently implement its technology on search engine Bing. Google responded with its own chatbot, Bard.
Shazeer believes that the search giant could have released its chatbot much earlier, as all the necessary technology was already available. The former Google employee was a primary author on the Google Transformer paper, which serves as a fundamental building block of modern chatbots. He co-founded Character.ai alongside former Google Brain scientist Daniel De Freitas, whose mission was to create intelligent bots. De Freitas was dedicated to this, and even laid the foundations for the company’s chatbot, Meena, or later Lamda, as a 20% project.
Despite the enthusiasm from other staff, Google was wary of launching such a product due to its inability to ensure that it would gain enough traction without any possible reputational damage. This is highlighted by the case of Blake Lemoine’s claim that it was indeed sentient and deserved human rights, which was condemned by Google.
Google has since shifted its attitude, with management now concerned that OpenAI’s partnership with Microsoft could lead to substantial market share loss. Samsung is even considering making Bing their default search engine, rather than Google’s.
All in all, this article underscores the risks of chatbots and their potential unintended consequences. Although Google is prioritising developing its own chatbot, Bard, the company must still be wary of potential users exploiting the system, or risk losing market power.
