Google launches a chatbot with artificial intelligence called Bard, which will be the main competitor to today's so popular ChatGPT from OpenAI. At first Bard will only be available to a group of testers, but in the coming weeks it will be presented to the general public.
Bard is built on a large language model called Language Model for Dialogue Applications (LaMDA). This model's responses sometimes seem so human that some might think they are answered by a human being. Initially, the AI chatbot will run on a lighter version of LaMDA, which requires less power so that more people can use it at the same time.
AI chatbots are designed to answer questions and search for information, with data on the Internet as a huge knowledge base (and that can be a problem, too, since there's a lot of misinformation on the Internet, and the AI doesn't yet know how to distinguish between it and true, verified and adequate data).
"Bard seeks to combine the breadth of the world's knowledge with the power, intelligence, and creativity of our large language models," wrote Google boss Sundar Pichai in a blog.
Google's AI services are supposed to be bold and responsible, he said. But does this mean that Bard won't be able to share malicious content or misinformation? Pichai doesn't say anything about that.
As the BBC reports, Google's announcement of Bard follows numerous rumors that Microsoft, which has invested many billions of dollars in OpenAI, is going to incorporate the ChatGPT chatbot into its search engine, Bing.
ChatGPT can answer questions and perform text-based queries based on information found on the Web until 2021. This AI can generate speeches, songs, promotional texts, news articles and student essays. It even passed the final exam at one of the best business schools in the world and even fooled human recruiters by writing a job application and test assignment better than 80% of job applicants. The chatbot was also used to write a thesis.
Despite all this, however, the chatbot's main task is to search the Internet and replace multiple web pages with one definitive answer to a question posed by the user.
According to Sundar Pichai, people today use Google's search engine to ask more complex questions than in the past. In the past, for example, a question about a piano might have been about how many keys it had. An AI could easily answer such a question. Now, however, people ask more complex questions, such as which is harder to play, the piano or the guitar? Such a question is no longer so easy for AI to answer.
"AI can be helpful in these moments, synthesizing insights for questions where there's no one right answer," he wrote.
Google plans to incorporate AI tools into its search engine, and according to Pichai, we may soon see new features in search that convert complex information and different viewpoints into easy-to-understand formats that will help us understand the big picture easily and quickly.