Google launched on Tuesday its Bard Artificial Intelligence tool, a competitor of ChatGPT, in public access, with the aim of improving the quality of its responses thanks to the increase in interactions with users.
The technology giant, owned by the Alphabet group, had announced the creation of Bard in early February, in response to the arrival, in November, of ChatGPT, developed by the OpenAI start-up in collaboration with Microsoft.
Capable of producing ’emails’, dissertations or lines of code with a simple request, ChatGPT is generating immense enthusiasm for generative artificial intelligence.
Bard’s use was initially limited to “trusted testers” before opening up to the general public on Tuesday.
However, the number of connections was restricted and a waiting list was created to manage the demand.
Access is currently only possible from the United States and the United Kingdom.
“As people start using Bard and test its capabilities, we will be surprised,” said Google leader Sundar Pichai in a message to his team and consulted by Agence France-Presse (AFP).
“Things will go wrong. But user feedback is essential to improve the product and the underlying technology,” he added.
The leader of the Californian group had been criticized internally for launching Bard too quickly to catch up with Microsoft.
The interface consists of a ‘site’, distinct from the Google search engine, with a space in which the user can type a question.
Asked by AFP what sets him apart from ChatGPT, Bard replied that, unlike his rival, he is “capable of accessing real-world information through the Google search engine.”
The chatbot [robô de conversação] He also highlighted that “it is still under development while ChatGPT is already available to the general public.”
“That means I’m constantly learning and improving, while ChatGPT will certainly remain unchanged,” he emphasized.
Already the vice presidents of Google, Sissie Hsiao and Eli Collins, highlighted in a publication on the group’s ‘site’ that they learned a lot by testing the Bard.
“The next big step to improve it is to get ‘feedback’ from more people,” they said.
“The more people use it, the better the large language models (LLM, a program that can generate answers to questions asked in everyday language) will be to better predict the answers that can be useful,” explain the two managers.
By receiving written data and conversations, the LLM algorithm can more accurately determine the relevant answer to a question.
Bard relies on LaMDA, a language model developed by Google to generate ‘chatbots’, the first version of which was launched by the Mountain View (California) group in 2021.
The two Google executives acknowledge that LLMs “aren’t perfect” and can “safely provide inaccurate, misleading or false information.”
Google indicates that it has implemented “safeguards” to contain the possibility of inaccurate or inappropriate responses, in particular by limiting the duration of exchanges in a dialogue between Bard and a user.
Since the launch of ChatGPT, several users have sought to push the ‘chatbot’ to the limit and have generated absurd, even worrying responses.
In an interview with The New York Times, Sissie Hsiao and Eli Collins said Google has yet to determine a business model and strategy for generating profit from Bard.
Source: TSF
