AI agents are not desirable in Vivaldi's browser.
In the ever-evolving digital landscape, one browser company is taking a different approach to artificial intelligence (AI) integration. Vivaldi, the browser developed by Jon von Tetzchner, has made it clear that it will not integrate Large Language Models into its browser, choosing instead to prioritise user control and discovery.
In a recent statement, Vivaldi expressed concerns about potential malicious actors tricking AI agents to access sensitive information such as emails or account details. This concern is not unfounded, as AI services, including those offered by major browser providers and some AI companies developing browsers, are vulnerable to potential attacks like prompt injections.
Last year, Vivaldi raised concerns about plagiarism, copyright infringement, and privacy violations with Large Language Models. The company believes that these models generate convincing-sounding lies, a threat that is becoming increasingly apparent with the rise of AI-powered chatbots.
Despite these concerns, other companies are pressing forward with AI integration. The Browser Company, for instance, is working on what is being called the first "real" AI browser as of June 2025. Their goal is to create a browser using AI agents, a trend that Vivaldi has chosen to avoid.
Vivaldi's commitment to user control and discovery is evident in their approach to browsing. The company believes that browsing should help people discover things, refine ideas, and make their own decisions. This philosophy is reflected in the videos on their website, which are designed to educate and inspire users.
In contrast to the hype surrounding AI in the browser, Vivaldi remains steadfast in its belief that the internet becomes less exciting without discoveries. The company consistently fights for a better web, a mission that sets it apart from those who are quick to embrace AI without considering its potential risks.
Recent reports have highlighted these risks. OpenAI's CEO, Sam Altman, has cautioned that ChatGPT's agent makes mistakes with no current solutions. Similarly, Anthropic has published a report on how cybercriminals are exploiting the AI chatbot Claude.
Despite these challenges, Vivaldi chooses people over hype. The company's goal is not to make users passive viewers but to empower them to explore, learn, and make informed decisions. In a world where technology is constantly evolving, Vivaldi's commitment to its principles is a refreshing reminder of the importance of putting people first.
