In brief
Bard is Google's AI service built as a response to OpenAI's ChatGPT. It runs on LaMDA, Google's large language model, and draws on information from the web to produce high-quality, fresh answers, including images, videos and other multimedia. Google launched it first in a lightweight version that needs less computing power, so it can scale to more users.
- Google tests Bard's answers by combining external feedback with internal testing so responses rest on real information and meet quality standards
- Google launched the Generative Language API, based on LaMDA, so other developers can build applications on its AI models
- Google plans to apply LaMDA, PaLM, Imagen and MusicLM to its products, starting with search, to answer more complex questions
Google ha annunciato il lancio di Bard, il suo nuovo servizio di intelligenza artificiale (AI) basato sul modello di lingua LaMDA, con l’obiettivo di offrire risposte di alta qualità e contenuti freschi e interessanti. Il nuovo servizio rappresenta la risposta di Google all’avanzamento di ChatGPT di OpenAI, il modello di generazione del linguaggio naturale tra i più avanzati oggi disponibili.
Come Google sta utilizzando l’AI
Google ha dedicato ampie risorse e investimenti all’AI, ritenendola la tecnologia più profonda e importante del nostro tempo, in grado di offrire opportunità per migliorare la vita di miliardi di persone e di sostenere la missione dell’azienda di organizzare le informazioni del mondo e renderle universalmente accessibili e utili.
Google ha fatto importanti progressi nell’AI, con il modello Transformer e il suo paper del 2017, nonché con i modelli di generazione dell’AI e i grandi modelli di lingua, che stanno attirando l’attenzione del pubblico. Con Bard, Google si propone di combinare la vastità della conoscenza del mondo con la potenza, l’intelligenza e la creatività dei suoi grandi modelli di lingua.
Bard di Google
Bard utilizza il modello di lingua LaMDA di Google per offrire risposte di alta qualità e contenuti freschi e interessanti. Il servizio si basa su informazioni raccolte dal web, e si propone di aiutare le persone a imparare cose nuove in modo creativo e curioso, come spiegare scoperte recenti della NASA a un bambino di 9 anni, o fornire informazioni sui migliori attaccanti nel calcio attuale, insieme a esercizi per migliorare le proprie capacità.
Il servizio è stato lanciato inizialmente in versione leggera, con la versione più grande prevista per il futuro, che richiederà meno potenza di calcolo e consentirà di scalare il servizio per raggiungere un numero maggiore di utenti. Google si impegna a garantire la qualità e la sicurezza delle risposte di Bard, combinando il feedback esterno con i test interni per assicurare che le risposte del servizio siano basate su informazioni reali e rispettino standard di alta qualità.
Il futuro dell’AI di Google
Google continuerà ad investire in tecnologie AI avanzate come LaMDA, PaLM, Imagen e MusicLM, in grado di offrire nuovi modi di interazione con informazioni di diversi tipi, dalla lingua alle immagini, al video e all’audio. Google si propone di utilizzare queste tecnologie per migliorare i suoi prodotti, iniziando con il motore di ricerca. L’obiettivo è rendere più facile per le persone comprendere il significato di ciò che cercano e trovare risposte anche alle domande più complesse.
Google sta lavorando anche per rendere più facile e scalabile l’accesso alle sue tecnologie AI più avanzate, attraverso la creazione di strumenti e API che consentano ad altre persone e aziende di costruire applicazioni.
How Bard works
Bard is powered by LaMDA, a large language model created by Google. Bard extracts information from the vast amount of data on the web, analyzes it, and creates high-quality answers in a creative and intelligent way.
The end result is a virtual assistant that can answer users’ questions efficiently, quickly, and accurately. Bard also offers interactive answers, not just text, but also images, videos, and other multimedia formats.
Google Search Evolves with AI
Google has a long history of using AI to improve search, and the company has introduced a number of cutting-edge technologies to do just that. For example, BERT, one of Google’s first transformer models, revolutionized understanding the subtleties of human language.
In recent years, Google has introduced technologies such as MUM, LaMDA, PaLM, Imagen and MusicLM, which allow you to access information in increasingly advanced and intuitive ways. By using AI, users can access multilingual content, images, videos and audio more effectively, and can also deepen their knowledge on a given topic through a variety of opinions and points of view.
AI for the development of new applications
Google is working to make AI accessible to everyone, not only through its own products, but also by creating tools and APIs that allow other developers to use its AI models to create innovative new applications.
The company has just launched its Generative Language API, which uses LaMDA as a foundation, and announced the creation of a series of tools and APIs that will allow creators to develop AI applications in a secure and scalable way.
In addition, Google is working with startups and other companies to provide them with the computing power needed to develop reliable and secure AI systems.
A promising future for AI
AI is an ever-evolving technology, and the future is bright for those interested in using it to improve their businesses and lifestyles.
Google is doing its part to pave the way for a future where AI is accessible to everyone and can be used to solve real problems and improve people’s quality of life.
Additionally, with continued research and development of ever more advanced AI technologies, the future is bright for those interested in developing innovative applications that use AI in ways never seen before.
Conclusions
In conclusion, Bard represents an important step for Google in the continued development of cutting-edge AI technologies. By combining the vast amount of data on the web with the pow
Frequently asked questions
What language model does Google Bard use?
Google Bard is powered by LaMDA, a large language model created by Google. Bard extracts information from data on the web, analyzes it and generates answers in a creative way. It was first released in a lightweight LaMDA version that requires less computing power, allowing Google to scale the service to more users.
What is the difference between Google Bard and ChatGPT?
Bard is Google's answer to OpenAI's ChatGPT. Both generate natural language answers, but Bard runs on Google's LaMDA model and draws on information gathered from the web to offer fresh content. Bard can also return interactive answers that go beyond text, including images, videos and other multimedia formats.
What can you ask Google Bard?
Bard is designed to help people learn new things with curiosity and creativity. Examples given by Google include explaining recent NASA discoveries to a 9 year old child, or giving information about the best strikers in current football together with exercises to improve one's own skills.
How does Google use AI in search?
Google has long used AI to improve search. BERT, one of its first transformer models, changed how search understands the subtleties of human language. Later technologies such as MUM, LaMDA, PaLM, Imagen and MusicLM help users reach multilingual content, images, video and audio, and explore a topic from multiple points of view.
Can developers build applications with Google's AI models?
Yes. Google launched the Generative Language API, built on LaMDA, and announced a series of tools and APIs that let creators develop AI applications securely and at scale. Google also works with startups and other companies to give them the computing power needed to build reliable and secure AI systems.
Last substantive revision: April 29, 2025.
