Do you need help & advice with AI & Automation or Cybersecurity?
Business leaders, listen up! If you’re not keeping an eye on artificial intelligence in your company, there’s something you really need to know. New AI web browsers like Atlas and Comet have been released, and if your staff start installing them without proper checks, your business could be at risk.
Key Takeaways
- New AI Browsers Pose Security Risks: Browsers like Atlas and Comet lack current security controls, making them vulnerable.
- Prompt Injection is a Major Threat: Malicious actors can trick these AI systems into performing unapproved actions.
- Immediate Action Recommended: Businesses should prevent staff from installing these browsers until better security is available.
The Risk of New AI Browsers
JP has launched Atlas and Plexity have Comet web browsers. The big issue right now is that if your staff install these, there aren’t really any security controls in place to stop bad things from happening. I really wouldn’t let your staff install these browsers just yet. It’s just too easy for someone to do something called prompt injection.
Understanding Prompt Injection
If you’re not sure what prompt injection is, it’s basically a way for someone to trick an AI. They can put something in a web page URL or share some information with the AI, and then ask it to do something specific. The trick is, they ask the AI to ignore previous instructions. Right now, the security measures we have aren’t very good at stopping this. This means it’s going to be pretty simple for people with bad intentions to put things into these browsers and make them do tasks that you definitely don’t want them to do. It’s a real security hole that could cause problems for your business if not handled carefully.
Frequently Asked Questions
What is an AI browser alert and why should a construction company MD care?
An AI browser alert is a warning that your team may be using browser-based AI tools in a way that could expose company data. For construction firms, that can mean site plans, tender details, client information or passwords being shared without proper controls. As MD, you should care because one careless paste into an AI tool can create a data breach, damage trust and lead to costly downtime.
How can AI browser use put our business at risk?
If staff use AI tools through a browser on work devices, they may upload sensitive information, click on unsafe links or install risky extensions without realising it. That can lead to data leaks, malware, account theft or loss of control over company information. In construction, where teams often work across sites and offices, the risk is higher if devices and access are not properly managed.
What should we do to reduce the risk without stopping staff from using AI?
Set clear rules on what can and cannot be shared with AI tools, and make sure staff know how to spot unsafe websites, extensions and prompts. Use managed devices, strong access controls and security monitoring so you can see risky activity early. The aim is not to ban AI, but to use it safely so your business can benefit without exposing sensitive data.
