In the age of agentic AI, convenience comes at a cost. Intelligent assistants like Comet, DeepSeek, and ChatGPT are rapidly transforming how we search, schedule, and stay organized. But as these tools become more deeply embedded in our digital lives, an uncomfortable question arises: what are we really giving up when we grant AI apps access to our personal data?
Recent concerns, highlighted by privacy advocates and technology analysts, reveal that the permissions we so casually accept may open the door to serious data privacy risks.
AI Is Asking for Too Much
Signal president Meredith Whittaker warns users to be wary of how AI apps gain access to personal data. She argues that some AI assistants are “asking to read your email, access your calendar, look at your photos, and see your contacts” requests that may seem harmless but could expose vast troves of sensitive information.
“You’re basically putting your brain in a jar and handing it over,” Whittaker told TechCrunch. And she’s not exaggerating. Once access is granted, many AI tools don’t just observe they learn, analyse, and even retain information for future interactions. This blurs the line between helpful personalization and invasive surveillance.
DeepSeek’s Meteoric Rise—and Its Red Flags
The China-based AI assistant made headlines last week after it surpassed one million global searches in under 48 hours. For a moment, it even overtook OpenAI’s ChatGPT as the most downloaded app on the iOS App Store.
But with this rise came a wave of scrutiny.
Within days of its spike in popularity, DeepSeek was pulled from app stores in several EU countries, including Germany, Ireland, and Italy, over serious GDPR concerns. Critics flagged the app’s aggressive data collection methods and lack of transparency over how user data is stored and processed. Even the U.S. Navy issued an internal memo banning its use on government-issued devices.
The Broader Risks of Agentic AI
The term “agentic AI” refers to tools that not only respond to queries but proactively act on your behalf. This means scheduling meetings, sending messages, or even making purchases often based on information pulled from your device.
To operate effectively, these systems need access to your calendar, contacts, files, and emails. But the risk lies in what happens after that access is granted. Some companies store and process this data in the cloud, where it may be exposed to third-party actors or retained longer than necessary.
And because many users skip reading privacy policies or permission prompts, they unknowingly give apps more access than intended.
How to Protect Your Personal Data
While AI can be an invaluable productivity tool, users should adopt a more cautious approach. Here are a few tips to regain control:
• Review App Permissions
Go through the settings of your AI tools and revoke access to data that isn’t essential.
• Use On-Device AI When Possible
Tools that process data locally, rather than in the cloud, offer better privacy protection.
• Be Mindful of New Apps
Especially those with unclear origins or poor documentation, like some rapidly viral tools.
• Watch for Regulatory News
Stay informed about app bans or investigations in your region these are often early warning signs.
AI Is Here to Stay But So Is the Need for Caution
Artificial intelligence will only become more capable and embedded in our lives. But that doesn’t mean we should sacrifice our privacy in the process. Tools like DeepSeek offer powerful functionality, but at what cost? And is that cost worth the convenience?
As Meredith Whittaker aptly put it, “just because a system can do something doesn’t mean it should.”
Before clicking “allow,” take a moment to consider what you’re really giving up. Because once your data is in the jar it’s hard to get it back.
AI apps like DeepSeek and Comet request deep access to personal data. Learn why experts urge users to think twice before granting permissions and how to protect your privacy.
