AI is rapidly shifting from something we interact with into something that can act on our behalf. We can ask AI to research, shop, plan, create and get things done for us. And increasingly, AI agents are connecting with the apps and services we already use – accessing information and learning our preferences to make those experiences more useful and personal.
And that shift is happening fast. New Gen telemetry shows nearly one in three desktop users are now interacting with agentic AI services.
That potential is exciting – but there’s an important tradeoff: The more AI can do for us, the more it may need to know about us.
Today, Gen is addressing that tradeoff with the launch of the AI Data Checker, a new tool within its Agent Trust Hub, designed to help people better understand how AI apps and agents may interact with their data.
AI is getting more personal. So is the data it uses.
AI agents and apps represent an important shift in our relationship with technology. As our chatbots have evolved to take on more of our complex tasks, they may connect with other apps, remember preferences and use personal context to deliver more personalized experiences.
While this information makes AI so useful, it also raises important questions.
What data does an AI service collect? What can it access? How might that information be used? Is it retained? Could it be used for training, personalization or advertising?
The answers may exist somewhere in privacy policies and terms of service, but finding and understanding them isn't always easy. We believe it should be.
Introducing the AI Data Checker
The AI Data Checker is designed to make your relationship with AI safer and more transparent.
As part of Gen's Agent Trust Hub, the tool helps people get greater visibility into how their data may be handled by the AI apps and agents they use.
The Agent Trust Hub is Gen's broader effort to help people navigate an increasingly agentic internet with greater visibility, understanding and protection. Since launching in February, it has expanded to include technologies such as VPN for Agents, purpose-built for autonomous agents, and Sage, Gen's consumer agentic security layer available through Norton AI Agent Protection in Norton 360. AI Data Checker adds another piece to that work: helping people understand the data relationship behind the AI tools they use.
The AI Data Checker builds on the tools already available through the Agent Trust Hub, which explores the data practices of popular AI services and helps translate complex policies into information people can more easily understand.
Using the AI Data Checker is simple: Enter your email address or username and the tool checks where that information may appear across AI services, helping you uncover which apps may have your data and understand what they can do with it. From there, you can explore individual services to see how they handle data — including whether information may be used to train AI, personalize advertising or be retained — and dig deeper into the policies behind those practices. The tool also scans for non-AI accounts where you may have connected and potentially shared personal information.
The Agent Trust Hub also lets people paste in the URL of an AI app or its terms page to scan the text and get a clearer picture of what they may be agreeing to.
The goal is simple: Help people understand what they're sharing with AI, how that information may be used and what choices they have.
People shouldn't need to be privacy experts to understand what happens to their information.
What some apps allow
The convenience of AI agents can come with data practices people may not expect.
Take Instinct, a personal AI assistant that can connect to other services and take actions on a user's behalf. According to an analysis of Instinct's terms, information shared with the assistant may be used to train AI models unless a user opts out or uses its protected Vault feature. And connecting another account can have implications beyond the connection itself: Instinct's terms state that even after a Connected Service is disconnected, previously indexed data from that service may still be used unless the user takes additional steps to request its deletion.
Then there is what an agent can do with that access. Instinct can make purchases on a user's behalf using payment information they've provided. Its terms make clear that those purchases are treated as purchases made by the user – and disputes are between the user and the relevant merchant.
Other AI services illustrate different kinds of data tradeoffs. Replika, for example, says it may collect information including location, contact information, identifiers and usage data. Apple's privacy disclosure for the app also says certain identifiers and usage data may be used to track users across apps and websites owned by other companies. Replika's own privacy policy says information about visits to its website may be shared with advertising partners for interest-based advertising, although conversations and content provided within the app are not shared with advertising partners.
These practices aren't all equivalent, and they don't necessarily mean a service is unsafe, but they may raise red flags for people who value their data and privacy.
Useful AI shouldn't require blind trust
Data itself isn't the enemy. AI agents need context to be useful. An agent helping organize your day isn't very helpful if it can't see your calendar. An agent helping plan a trip may need to know where you're going and your travel preferences. An agent helping manage your inbox needs access to your messages.
The answer isn't AI that knows nothing about us. It's giving people greater visibility, understanding and control over what they choose to share.
People should be able to make informed decisions about which AI services they trust, what information they give them and how that information is used.
At Gen, we believe people should be able to embrace what AI makes possible while remaining informed and in control of their digital lives.
The more AI knows about us, the more important it becomes for us to understand what it knows – and what happens to that information. That visibility can help people use AI with greater confidence and live fearlessly in an increasingly agentic world.
