Brave Launches AI Assistant: Leo

The well-known browser developer Brave has released its native AI assistant Leo, taking an important step in protecting user privacy and improving the accuracy of AI interactions, and it is released together with the desktop version 1.6. Leo is powered by Meta Llama 2 as its underlying model, which can respond to user queries based on the web content they visit, effectively addressing issues related to AI-generated content.

Leo is an extension of the Brave search AI summary tool launched earlier this year, which can be accessed directly from the search bar. During the testing phase in August, thousands of developers and users downloaded and evaluated the browser and Leo through the Nightly channel (version 1.59), and it was finally integrated into Brave's 1.60 version.

One unique feature of Leo is its commitment to user privacy. Unlike other chatbots, Leo does not collect conversations, track users, or generate responses out of thin air. Instead, it relies entirely on web content to provide accurate and relevant information.

The free version of Leo is based on the highly secure Llama 2 model, which is a specialized variant of the Meta open-source model. However, Brave has also introduced Leo Premium, a paid service that costs $15 per month. Leo Premium is equipped with the Claude Instant model developed by Anthropic, which emphasizes logical reasoning and code writing. This model provides more structured responses, enhanced instruction execution capabilities, and better abilities in mathematics, programming, multilingualism, and question-and-answer interactions.

To further improve the accuracy of responses, Brave has integrated Anthropic technology and used Brave's search API to train the latest Claude 2 model. This approach helps Claude products achieve Retrieval-Augmented Generation (RAG), resulting in more accurate responses and reducing the illusionary tendencies of generative AI.

In terms of security and privacy, Brave has taken extensive measures. In the free version, Leo's conversations remain anonymous and confidential, with no recording of interaction content. The data is not used for model training, and no account or login is required. Reverse proxy technology ensures that all calls are made through anonymous servers, preventing Brave from establishing any association between calls and user IP addresses.

For users who choose to use Leo Premium, an untraceable token is issued during registration to ensure the security of the subscription verification process. This means that Brave cannot associate usage activity with user purchase information, ensuring complete privacy. In addition, users' email addresses are only used for subscription verification and cannot be traced.

Looking ahead, Brave plans to introduce additional models in the Premium version and make improvements in network speed limitations, conversation quality, and exclusive features for subscribers.

Currently, Leo and Leo Premium are available in Brave 1.6 desktop version, and they are planned to be launched on Android and iOS platforms in the coming months. This innovative development marks significant progress in browser technology and AI integration, reaffirming Brave's commitment to user-centric innovation with a focus on privacy protection.