Brave unveils Leo, its privacy-respecting AI assistant, a major breakthrough

Publié le 23 February 2025 à 10h32
modifié le 23 February 2025 à 10h32

After three months of trials and tests, Leo, a virtual assistant powered by artificial intelligence (AI) for the Brave web browser, is officially launched. This chatbot stands out from others by ensuring the respect of your privacy. Available in version 1.60 of the browser, Leo is accessible for free in its basic version. The latter uses Meta’s Llama 2 language model.

The Leo chatbot allows you to summarize web pages viewed and videos watched in real time. Brave takes great care to ensure the privacy of its users. Thus, Leo does not retain conversations and does not use them to train other models, without requiring an account or login to use their services, as explained by Brave in their statement.

Enhanced data protection through a reverse proxy

To ensure this confidentiality, all requests sent to the chatbot are routed through a reverse proxy. This prevents any association between the information provided and the users’ IP addresses. The developers of the chatbot have designed an AI that safeguards privacy while offering advanced features.

A paid Premium version with other language models

For those wishing to enjoy a more advanced version of Leo, Brave offers a Premium version with a monthly subscription. For about fifteen dollars per month, users can access other language models, including Claude Instant developed by Anthropic.

Faster and lighter than other models, Claude proves to be more efficient in logical reasoning and programming according to Brave. The company adds that other language models will be integrated into the Premium version of Leo.

An AI assistant compatible on desktop, Android, and iOS

This innovative chatbot is now available on the Brave browser for computers. Like other AI assistants on the market, Leo allows users to request various tasks, such as creating a summary of a web page or video, translating or rephrasing pages, and even generating new content. In their announcement, Brave Software emphasizes the preservation of user privacy and insists that conversations with Leo are not retained on their servers.

Responses provided by the assistant are deleted immediately after consultation and are not used to train other models. For smartphone and tablet enthusiasts, Leo will be available on Android and iOS systems in the coming months.

Brave facing criticism and accusations

Last July, Brave was criticized for allegedly selling copyrighted information to train its AI models without consent. Josep M. Pujol, head of research at Brave, responded to these allegations by stating that “Brave Search has the right to monetize and set usage rules for its search engine launch,” adding that this is part of the standard features expected in all search engines.

A gradual rollout of Leo

The Leo chatbot will be gradually rolled out on desktops over the coming days. As for Android and iOS mobiles, the assistant is expected to be available in a few months. Interested users can download the development version on Brave’s official website. However, caution is advised as these versions are primarily intended for developers and experienced users and may contain bugs.

Leo thus represents a new step in the continuous improvement of web browsers, offering their users innovative features while considering the protection of their privacy.

actu.iaNon classéBrave unveils Leo, its privacy-respecting AI assistant, a major breakthrough

Plan your tasks with ease: an AI agent to manage your meetings, errands, and flight reservations

optimisez votre emploi du temps grâce à notre agent ia intelligent. planifiez vos réunions, gérez vos courses et réservez vos vols en toute simplicité. libérez votre esprit et concentrez-vous sur l'essentiel avec une assistance technologique à la pointe!

The historical videos generated by AI spark debate: educational tool or source of misinformation?

découvrez comment les vidéos historiques créées par l'intelligence artificielle soulèvent des questions essentielles : sont-elles un véritable outil pédagogique ou une potentielle source de désinformation ? analysez les enjeux et les perspectives d'une technologie en plein essor.

Grok 3: Elon Musk’s artificial intelligence makes a blunder live during its unveiling

découvrez comment grok 3, l'intelligence artificielle développée par elon musk, a fait des erreurs surprenantes en direct lors de son lancement. analyse des implications de ces faux pas et des réactions du public.

OpenAI reaches 400 million weekly users and aims for an unprecedented valuation

découvrez comment openai a atteint 400 millions d'utilisateurs hebdomadaires et explorez ses ambitions pour atteindre une valorisation inédite, redéfinissant ainsi le paysage technologique.
plongez dans l'univers fascinant de l'architecte derrière les coulisses du budget français. découvrez comment une seule entité controle les ressources financières et influence les décisions qui pourraient façonner votre avenir. ne laissez pas passer cette analyse approfondie sur le pouvoir, l'argent et l'impact sur votre quotidien.

Intelligent Artificial: the 10 most efficient models to watch in February 2025

découvrez les 10 modèles d'intelligence artificielle les plus prometteurs à suivre en février 2025. cet article vous présente des innovations marquantes qui redéfinissent le paysage technologique et vous aide à rester à la pointe des tendances ia.