- Technology What is Bard and how Google's artificial intelligence works
- Technology Bard, Google's artificial intelligence, learns Bengali on its own
Bard, Google's artificial intelligence, now available in the U.S. UU and United Kingdom since March of this year, finally arrives in our country and Jack Krawczyk, Director of Product Management of Google, has come forward to tell us some of the improvements he has implemented since March and how it has adjusted to the legal landscape of the European Union.
This artificial intelligence, which is presented rather as a new function fully integrated into the Google ecosystem, works as a bard of the digital age (hence its name, according to Google). This skilled reciter of stories in digital version, is able to captivate and surprise with creative and original stories, in a versatile way, regardless of the subject to be treated and adapting its tone to the context required by the user. Even if it requires some human 'empathy'.
What this technology offers is a way to clarify some basic ideas and develop them creatively, wanting to move Bard away from the concept of Chatbot that has been assimilated with respect to other systems seen so far, such as ChatGPT, allowing not only to interact with a generative text chat, also with a voice chat totally in Spanish, making the experience much more organic.
Giant steps away
Although Bard does not resonate as much as ChatGPT it is clear that it comes to compete with it. In fact, there is a whole war to become the owner of the most efficient AI, a battle that Elon Musk has recently joined with. Google Bard also has the ballots
to become one of the most used AI, since it has the advantage of integration with the Google ecosystem.
Little by little, Bard is adding languages. For now it has more than 40, including Spanish, which has been presented this Thursday, and with plans to introduce Catalan in the future. In addition to having access throughout Europe and Brazil. This race, far from being in a hurry and without looking back, has been planned from stage to stage, so that when asked why Spanish was not included at the beginning and in Spain, being one of the first five languages in the world in number of speakers or in its first update, in which Japanese and Korean were added.
Jack Krawczyk replies that there have been two important reasons why it has been made to wait, one for the language and one for the territory. The first is due to the interaction with the user, since from Google they want Bard not only to give answers that can be understood in the language. They aim to adapt to the required circumstances and specific contexts, from a formal letter addressed to the boss, to a children's story, developing in depth the voice options, since hearing something out loud can help you focus the ideas in another way.
Commitment to transparency
The second and most important reason is due to the privacy policies of each territory. In this way, they wanted to ensure total legal transparency and with the user, taking care not to have problems, as already happened to ChatGPT in Italy. Another very important asset, in what goes forward the Law of Artificial Intelligence. Since those who do not conform to future regulations, will see their area of action limited. That is, they would face having services vetoed in the European Union, something that Google is trying to take care of.
To ensure the safety of users, in addition to having launched a new Bard Privacy Help Center to facilitate the review of privacy controls available to users. Now, this tool has several privacy options available to the user, such as customizing how long you want Bard to store your data in the personal Google account, which by default is 18 months, being able to modify or deactivate and thus delete that history immediately. As well as the option to pin, rename and resume conversations easily, or share them with other people.
Little by little Bard is integrating with different tools of the Google ecosystem, such as Docs, Lens or Colab, allowing functions such as showing images in their answers, in addition to their respective text and voice descriptions. As a novelty, you can recognize images through Google Lens and offer assistance regarding what is asked about that same image.
This way you can help get a detailed and personalized description of a landscape and share it with someone who cannot enjoy the image on their own. So Bard can become a very useful tool for people with visual problems. Another new feature is to allow you to export Python code created in Google Colab that is generated with Bard, or even ask it to explain how it works. It also allows you to share conversations with Bard on social networks or with specific contacts. A range of possibilities that is still beginning to open up and whose competition is voracious.
According to the criteria of The Trust Project