SHARE
The risks of AI tools are well known, but recent reports reveal exactly how scammers steal sensitive data, passwords, files, and more. Knowing these tactics is key to staying safe and avoiding major security breaches. Let’s break down the restrictions you need to follow.
AI browsers can seem convenient, but they carry hidden dangers. Research by Brave has shown that AI browsers, such as Comet by Perplexity, can be vulnerable to prompt injection attacks. For example, AI agents in these browsers can be easily tricked into exposing confidential information. This allows malicious actors to manipulate the AI agent to gain access to sensitive data, such as passwords or files.
To stay safe, avoid AI-powered browsers like Comet by Perplexity for accessing sensitive data or performing critical tasks. Even if some vulnerabilities are patched, new risks keep emerging. AI browsers evolve fast, and loopholes appear quicker than security teams can fix them.
Team members must not connect any company accounts to AI chatbots, such as Claude, ChatGPT, or others. Integrating AI chatbots with other services can lead to sensitive data leaks. Specifically, a study presented at the Black Hat conference showed that vulnerabilities in ChatGPT integrations could allow sensitive information from Google Drive to be extracted without the user’s involvement.
To minimize risk, limit the use of:
Hackers can gain full access to email and cloud data through these integrations. Even one unsafe connection can compromise the company’s internal information. Following these rules strictly is essential to protect sensitive data.
Google Workspace integrations (e.g., Drive or Workspace) highlight how AI and cloud security issues can expose sensitive files. At Black Hat 2025, researchers demonstrated how a single poisoned document in Drive could trigger data leaks. Cybersecurity company Tenable also confirmed vulnerabilities that attackers could exploit to access emails and internal files.
To reduce risks, do not connect AI to work accounts (Gmail, Drive, Workspace). Use it only with separate test profiles until Google strengthens its security measures.
AI technologies are everywhere, and we use them daily to improve our work and life. They make tasks faster, smarter, and more efficient. But while enjoying these benefits, we must never forget the risks. Staying cautious, informed, and following safe AI usage guidelines is essential to protect sensitive data and use AI safely.