Since the American giant saw the chatbot as a rival to its search engine, it would have mobilized several teams to develop products and deal with this “threat”.
OpenAI’s ChatGPT chatbot was available to the public for a few weeks and was an instant success. Based on the GPT-3 language model, it is capable of conversing with people, answering all kinds of questions or solving complex problems with surprising speed and precision. Although this chatbot is not yet perfect, it worries online search giant Google.
Since the American dominates the search engine market, he sees ChatGPT as a threat. According to New York Times, he triggered a “code red”, fearing that this chatbot could later replace his search engine. Its CEO Sundar Pichai reportedly attended several meetings to redefine the company’s artificial intelligence (AI) strategy, and several teams were reportedly redeployed to help develop AI prototypes and products. Google could reveal the results of this work next May at its annual developer conference, Google I/O.
Risks for Google
The Mountain View company has been working on chatbots for several years. The technology at the heart of ChatGPT also comes from its labs, as indicated by the US daily. In addition, the company already has a chatbot similar to Open AI. Called LaMDA, it was the subject of particular attention this summer after a former Google engineer claimed that this AI was equipped with sentience. Currently, the company only allows certain people to test it, a way for it to avoid problems.
In fact, if this chatbot is only available as a demo, it’s for a good reason. These chatbots are trained on huge amounts of data published on the internet. They can thus provide quality information, just as they can generate problematic answers. Meta’s BlenderBot3 chatbot, for example, went off the rails days after its launch this summer, making false claims based on its interactions with people online. This type of issue can damage the reputation of a large company like Google, while smaller companies like OpenAI can take the risk of making their chatbots publicly available, especially if it allows them to grow and develop.
Google may also be reluctant to implement this type of technology in terms of its business model. The majority of its revenue comes from advertising. However, chatbots are not suitable for serving ads, unlike their search engine counterparts. Concretely, using a chatbot to ask a question rather than a search engine means that fewer people click on ads.