Do you need help & advice with AI & Automation or Cybersecurity?
AI at work promises to make life easier, but sometimes new tools show up before they’re safe to use. That’s exactly what’s happening with the recent release of ChatGPT’s Atlas and Perplexity’s Comet browsers. They look handy, but for now, keeping them out of your business is the smartest move.
Key Takeaways
- Atlas and Comet browsers are not yet secure for business use.
- Easy-to-miss threats, like prompt injection attacks, can put your organisation at risk.
- Productivity boosts aren’t worth the potential headaches from a data breach or worse.
What’s New with AI Browsers?
Atlas (from ChatGPT/Chip) and Comet (from Perplexity) have launched with much excitement around how they can help staff work faster. They’re special web browsers powered by artificial intelligence—they read web pages, summarise info, and sometimes even automate tasks for you.
Sounds great. Except, right now, these tools don’t have proper controls built-in to keep sensitive information safe. They’re easy to install, and staff might be tempted to give them a go even though they haven’t been tested for business-level safety.
Why Atlas and Comet Aren’t Ready for Business
The main concern is something called prompt injection. Here’s how it works:
- Someone puts hidden code or sneaky instructions inside a website or a bit of shared data.
- The AI browser reads it as a normal part of the page, but it can actually tell the AI to do things you definitely don’t want—like sharing business secrets, skipping security steps, or ignoring your company policies.
- Because the controls in Atlas and Comet are weak at the moment, there’s a real risk here.
If you’re running a company, you’re looking after people’s data, maybe client info, and your own secrets. The last thing you want is some new tech going rogue because it can’t tell a trick from a regular instruction.
Here’s a look at how prompt injection plays out:
| Threat | What Could Happen |
|---|---|
| Data Leak | Sensitive info sent out by mistake |
| Bad Automation | Unwanted tasks or emails sent |
| Policy Ignore | AI skips company rules |
The Illusion of a Productivity Boost
It’s easy to get excited about being more productive. These AI browsers might save a little time—but when you weigh that against a possible data breach or all the effort needed to clean up after a mishap, it’s just not worth the risk right now.
Letting staff install these tools might look like you’re ahead of the curve. But if the controls aren’t there (and they’re not), it’s your business on the line if something goes wrong.
What Should You Do Next?
It’s time to set some strong ground rules:
- Block Atlas and Comet at work. Don’t let staff install or use these browsers—at least until better security gets built in.
- Talk to your team. Let people know why these tools are off-limits for now.
- Update your AI policies. If you don’t have one yet, it’s a good time to get a simple policy in place.
Even if you’re a small business, a basic AI policy can make a big difference. There are free templates available if you need something to get started. No shame in not knowing where to begin—this stuff is all brand new for everyone.
Real World Thoughts—It’s Just Not Worth the Risk Yet
I’ve tried playing with the Comet browser myself. Parts of it are clever, sure, but for business? It’s not ready. The controls just aren’t tight enough. It’s tempting to let your team try the latest thing, but sometimes you need to wait for the tech to mature. No one wants to be the headline for the next big breach.
At the end of the day, protecting your organisation should always come first. If you’re not certain, don’t risk it. Your business’s safety is more important than being first to use the newest app.
