Jump to a section
Connecting AI to your apps is powerful, but anything it reads, including emails and invites from strangers, could contain hidden instructions. Connect only what you need, check before it sends or deletes anything, keep it away from banking and passwords, and disconnect what you're not using.
What connecting unlocks
Once an assistant can see your apps, it can do the work rather than just talk about it. Google says Gemini can connect to other apps to, for example, summarise emails from Gmail or create events in Google Calendar. On the web, you manage these under Customize, then Connectors, where you can add an app or disconnect it.
The risk most people haven't heard of
It's called prompt injection. The UK's National Cyber Security Centre describes it as "when an attacker creates an input designed to make the model behave in an unintended way", for example to reveal confidential information.
When an AI reads your inbox, it reads emails from strangers too. If one of those contains hidden instructions, the AI might follow them. Privacy International gives the example of a malicious calendar invite leading to the disclosure of emails, and argues that the industry has no real answers to the risk yet.
It also notes that connectors typically need full access to the service they connect to, so a connected assistant can reach everything in it, not just the bit you had in mind.
NCSC: AI and cyber security Privacy International: risks in turning chatbots into agents
Is it worth it?
For many people, yes, as long as you're deliberate about it. Summarising a busy inbox or turning a school email into calendar entries saves real time. The aim isn't to avoid connecting apps; it's to connect them with your eyes open, and to keep a human check on anything that can't be undone.
AI chatbots and your privacy: what they keep, who sees it, and how to take control
How AI chatbots actually work, and why it changes how you use them
Five habits for connecting AI to your apps
- Connect one app at a time, and only the ones you'll actually use
- Check what the AI is about to do before it sends, books or deletes anything
- Keep it away from banking, password managers and anything you couldn't undo
- Be wary when it's reading content from strangers: emails, invites, web pages
- Review connected apps every few months and disconnect what you don't use
Sources
These sources support the guide. Tool settings, prices and policies change, so dated source boxes show when those details were checked.
- UK National Cyber Security Centre: AI and cyber security, what you need to know (13 February 2024)
- Privacy International: risks in turning AI chatbots into AI agents (18 September 2025)
- Google: use connected apps with Gemini Apps
AI helps with research and drafting. Dan checks the sources and is responsible for what is published.
How we check