- Gartner warned organizations to block AI browsers
- AI browsers can threaten data, leak financial information and transmit identifying information
- Employees could use AI browsers for cybersecurity training
Analyst firm Gartner has advised organizations to prevent the use of AI browsers, warning of the risk of data exposure or autonomous actions performed by agent browsers on malicious websites.
The key observation made by VP Researcher Dennis Xu, Senior Director Analyst Evgeny Mirolyubov, and VP Analyst John Watts was that “AI browser defaults prioritize user experience over security.”
AI browsers, such as OpenAI’s ChatGPT Atlas, are often used to improve efficiency using autonomous navigation, workflows and data collection, but they can be tricked by malicious web pages into collecting and transferring sensitive information such as bank account details, credentials and emails.
“Happy to hack you 🙂”
Gartner analysts have summarized an Agent Browser to include two key features:
- The ability to interact with web content using the developer’s AI model, providing functions such as content summarization, data collection, translation and search capabilities.
- The ability to perform tasks autonomously on websites, including within authenticated sessions.
Many agent browsers do not allow the use of AI functions within a local LLM, meaning that user data, ranging from web content to browsing history and open tabs, “is often sent to the cloud-based AI backend, increasing the risk of data exposure unless security and privacy settings are deliberately strengthened and centrally managed.”
Ultimately, it is up to each organization to perform an assessment of the AI browser back-end services to see if they comply with the organization’s cybersecurity and data protection policies. But even if they are adopted, they can still be used in ways that pose additional risks to the organization.
In this case, the user himself can provide the browser with an unnecessary amount of sensitive information simply by opening the sensitive data in the same web browser window while using the browser’s AI assistant.
Additionally, because browser agents can perform actions autonomously, Gartner warns that employees may be “tempted to use AI browsers and automate some mandatory, repetitive, and less interesting tasks,” such as cybersecurity training.
Gartner suggests that organizations that continue to use agent browsers should “inform users that anything they view could potentially be sent to the AI service’s backend to ensure that they do not have highly sensitive data active on the browser tab when they use the AI browser sidebar to summarize or perform other autonomous actions.” »
Commenting on the Gartner advisory, Javvad Malik, principal security awareness advocate at KnowBe4, said:
“AI capabilities have introduced tensions into cybersecurity, forcing people to weigh the tradeoff between productivity and security risks. While agent browsers promise many features to improve the user experience, we are still in the early stages where the risks are not well understood and the default configurations prioritize convenience over security, something we see in many technologies.
“However, blanket bans are rarely sustainable long-term strategies. Instead, the focus should be on risk assessments that evaluate the specific AI services that power these browsers. This can allow for measured adoption while maintaining necessary oversight. As we see more AI agents making their way into all aspects of technology, organizations must put playbooks in place to evaluate and protect AI agents, and enable them to work within the organization according to their own needs and appetite for risk.”
Via The register
The best business laptops for every budget
