AI browsers like Perplexity Comet and OpenAI’s ChatGPT Atlas are gaining popularity in workplaces, but experts warn they pose serious security risks. According to Gartner, companies should block these tools for the foreseeable future. The main concern is that these browsers can expose sensitive data and introduce new cybersecurity threats that organizations aren’t ready to handle.

The Growing Use of AI Browsers in Businesses

Despite the risks, many companies are already adopting AI browsers. Recent data shows that about 27.7% of organizations have at least one employee using Atlas, with some seeing up to 10% of staff actively using it. The technology sector leads the way, with 67% adoption, followed by pharmaceuticals at 50% and finance at 40%. The launch of Atlas on October 21 sparked a surge in downloads, with it being downloaded 62 times more than Perplexity Comet, which was released earlier in July.

This increased interest has led to a spike in overall AI browser downloads, especially after Atlas’s debut. Cybersecurity experts warn that this rapid growth is happening without proper safeguards in place, raising alarms about potential vulnerabilities and data leaks. The trend underscores how quickly these tools are becoming embedded in enterprise workflows, often without comprehensive security measures.

The Security Risks and Data Privacy Concerns

The main worry with AI browsers is how they handle data. When these tools send browsing history, open tabs, or active web content to the cloud for analysis, companies lose control over their information. For example, Perplexity’s documentation states that Comet may process some local data using its servers, meaning it can read the contents of emails or other sensitive documents to complete tasks.

Experts highlight that once sensitive data is sent to an AI service, it’s often impossible to fully recover or trace. This creates a risk of irreversible data loss. Additionally, AI browsers can perform actions autonomously, like filling out forms or making transactions. This ability raises concerns about mistakes or malicious activities that could result from errors or deception, such as prompt injections or phishing attacks.

Traditional security controls aren’t enough to protect against these new threats. AI browsers can navigate websites independently, which means they could be tricked into visiting malicious sites or executing unauthorized transactions. If an AI browser is deceived into performing harmful actions, it could lead to significant security breaches, credential theft, or data exposure. Experts advise organizations to be cautious and consider blocking these tools until better safeguards are available.

Overall, while AI browsers offer new possibilities for productivity, the current security landscape makes them risky for enterprise use. Companies should prioritize understanding these risks and implementing measures to protect sensitive information. Until the technology matures and robust security controls are introduced, the safest approach is to keep AI browsers out of organizational networks.

Inspired by

• https://www.computerworld.com/article/4102569/keep-ai-browsers-out-of-your-enterprise-warns-gartner.html

Sources

• cyberhaven.com
• perplexity.ai
• teamwin.in
• layerxsecurity.com