When the artificial intelligence race seemed to have slowed down, Bard, Google’s generative artificial intelligence (AI) chatbot, arrives in Spain to step on the accelerator again. The tool —which was already working in more than 180 countries after a trial period of several months— is available as of this Thursday in the 27 countries of the European Union and in Brazil; and in 40 new languages, including Spanish, Chinese, Arabic and Hindi. This is the biggest expansion since the technology was introduced in February in response to ChatGPT, the OpenAI product whose success has forced the biggest tech companies to launch their own intelligent chatbots. “Bard is a tool to increase the imagination. He is a creative collaborator, a direct access to a generative artificial intelligence model driven by the most powerful computer in the world, the human brain”, explained the director of product management at Google, Jack Krawczyk, during a press conference held yesterday Wednesday to announce the arrival of the product in Spain and its new functions.
The tool has your own URL and user interface, is free and is presented as a complement to Google search, and not as a search engine itself. At the moment, the chatbot is only available in Spanish, although Krawczyk has assured that work is being done to make the other languages of the peninsula also available, such as Catalan, Basque and Galician: “Right now Bard is capable of responding in these languages, but we need to train him to answer responsibly.” In addition to writing, the chatbot is also capable of speaking in Spanish and reading responses aloud after clicking on a speaker icon, much like what can already be done with Translate, Google’s popular translator.
This is not the first experiment with artificial intelligence carried out by the company, which in fact is one of the companies that invests the most time and money in research in this field. The Google Brain division, in collaboration with the British company DeepMind, acquired in 2014, is at the global pinnacle of artificial intelligence. the research project transformer and his pioneering article, presented in 2017, served as the main pillar on which the so-called advanced generative artificial intelligence has been forged.
Last summer, Google developed several projects related to AI, such as the Dall-E2 or Midjourney text-based image generators, which caused some controversy over copyright violation and the danger of taking work away from illustrators. However, only after the boom generated by the arrival of ChatGPT —the bot conversational platform of the Open AI company capable of answering users’ questions quickly and apparently correctly—, the bigtech have decided to speed up the development of similar tools.
What can be done with Bard
Like ChatGPT and other AI chatbots, Bard relies on large language models (large language models, LLM), a type of technology that is capable of learning and generating answers through the analysis of large databases on the Internet. Therefore, Bard can also make mistakes or make things up, and it’s never a good idea to use him without checking your answers.
It is not intended to replace Google Search, the company’s search engine. In fact, it connects to Google’s website index so that its responses can include the most recent information published on the Internet, although the sources it includes are not always the most reliable. It is intended as a complement, one more tool to generate ideas, invent a more or less long story, give advice on destinations for the next summer trip or summarize a text, and even an audio. “I often wake up in the morning with too many ideas in my head. What I do is speak out loud for a few minutes and then ask Bard to put my ideas in order, and to summarize them in three fundamental points”, the Google executive gives as an example. The tool features five different options to change the tone and style of responses: easy, long, short, professional, or informal.
One of the biggest differences from ChatGPT—besides the ability to play responses aloud, an option the OpenAI tool doesn’t have—is the ability to export and share files from other company products, such as a Gmail email or a Google Docs document. Among the new functions, there is also the option to pin and rename conversations, have several queries at the same time and resume a conversation started at another time.
Another of the tools that complement Bard is Google Lens, which allows you to upload images to obtain information about it —such as identifying what it represents or writing a description— or asking the chatbot to generate a text from it, although this function is currently a novelty available only in English.
Privacy Concern
Bard’s delay in reaching Spain and the other countries of the European Union is due in part to the meteoric and uncontrolled career of ChatGPT, the product with the fastest growth in history. This unprecedented success has caught the attention of the European authorities, who have shown their concern, especially in what has to do with the protection of user data. In June, the European Parliament approved negotiating the first AI law in the world, with the idea of becoming, by 2026 at the latest, the first region to regulate this technology, which, in addition to its great promises, is also frightening because of the possibility of changing society.
“In a region like Europe, where there are privacy regulations, we do not believe that the correct way to do it is to ask for forgiveness after the fact. We think that it is better to talk with the regulators to be sure that our principles are aligned”, explains Krawczyk. When users interact with Bard, Google collects data such as conversations, location, comments, and usage information. Users can choose how long they want Bard to keep their data: By default, Google stores Bard activity for a maximum of 18 months, but this can be changed to three or 36 months, or turned off entirely.
You can follow THE COUNTRY Technology in Facebook and Twitter or sign up here to receive our weekly newsletter.
