DOES AI TRACK OLDER ADULTS? WHAT CAN YOU ACTUALLY CONTROL?

Introduction

When people ask if AI tracks them, they're usually worried about two different things: whether AI companies are recording their conversations and storing personal information, and/or whether AI is monitoring their broader online activity like some sort of Orwellian surveillance system.

The answer to the first question is mostly yes (with some important caveats), and the answer to the second is mostly no (although it's more complicated than that. Surprise!).

What AI companies actually collect

When you use an AI service like ChatGPT, Claude, or similar tools, the company running that service collects certain information by default, and understanding what they collect helps you make informed decisions about what to share.

Your conversations are the most obvious thing being collected because when you type messages to an AI, those messages go to the company's servers where they're processed to generate responses. Most companies claim they don't use your conversations to train future AI models unless you explicitly opt in, but they do temporarily store them to provide the service.

Some companies delete conversations after a certain period, others keep them indefinitely (or until you manually delete them), and the specifics vary by company and by which version of the service you're using. Free versions typically have looser data retention policies than paid enterprise versions.

Account information gets collected if you create an account, which might include your email address, payment information if you're using a paid service, and basic usage data like when you logged in and how often you use the service.

Technical metadata (data about data) is collected automatically, including your IP address (which reveals your approximate location), what device and browser you're using, and how you interact with the service. This is standard for most online services and isn't unique to AI.

What AI companies don't collect (usually)

AI services don't typically track your broader internet activity across other websites unless you're using browser extensions or plugins that integrate AI into your web browsing. If you go to ChatGPT's website, have a conversation, and then leave, ChatGPT doesn't know what else you're doing online.

However, if you install browser extensions (small programs that add features to your web browser), that changes things. For example, if you install a "ChatGPT sidebar" extension that lets you use ChatGPT on any webpage, that extension needs permission to see what websites you visit and what content is on those pages. The extension might send information about the websites you're browsing back to the AI service to provide its features.

Similarly, if you use AI-powered writing assistants like Grammarly or browser extensions that summarize web pages using AI, those tools can potentially see everything you type or every page you visit, depending on what permissions you granted when you installed them. This doesn't mean they're definitely tracking you, but they have the technical capability to do so if their privacy policy allows it.

The key difference is: visiting an AI website means only what you do on that site can be seen and maybe retained. However, installing an AI browser extension could mean that your activity could potentially be seen across all websites, depending on what permissions the extension requires.

AI doesn't access your files, photos, emails, or other data on your device unless you explicitly upload or share them during a conversation. If you paste text from a document into an AI chat, the AI sees that text, but it doesn't have access to other documents on your computer.

AI doesn't listen to your conversations through your microphone or watch you through your camera unless you're using specific features that require those permissions (like voice input or video calls), and even then it only has access when those features are actively being used.

The difference between tracking and data collection

There's an important distinction here that gets confused in conversations about privacy.

Data collection means the company receives and stores information you provide directly (your messages, your account details) or that which was generated as part of using the service (metadata about when you logged in, what device you used). This is not only what AI companies do, but almost all other companies behind a website, one way or another.

Tracking typically means following your activity across multiple websites or services to build a profile of your interests and behaviour, often for advertising purposes. This is what companies like Google and Facebook do extensively, but it's not what most AI chatbots do because they're not advertising platforms.

That said, some AI services are owned by companies that do track users across their other products. If you use Google's AI tools while signed into your Google account, Google can potentially connect that AI usage to your broader profile, though they claim they keep certain data separate.

What you can control

You have more control over AI data collection than you might think, though exercising that control requires being intentional about it.

Don't create an account if you're concerned about data retention because some AI services let you use them without signing up, which limits how much personal information they can tie to your usage. You'll lose features like conversation history, but you'll also limit what data is stored.

Delete conversations regularly if the service allows it, which many do through account settings. This doesn't guarantee the data is immediately purged from all backups, but it removes it from active storage and your visible history.

Use private or incognito browsing mode when using AI services because while this doesn't prevent the AI company from collecting data during your session, it does limits what information your browser shares and prevents the service from accessing cookies from your other browsing.

Read privacy policies (I know, boring) because they explain what data is collected, how long it's kept, and what options you have for deletion or opting out of certain uses. The policies are tedious, but they contain important information about what you're agreeing to.

Consider paid versions for sensitive work because enterprise or paid subscriptions typically offer stronger privacy guarantees, clearer data retention policies, and sometimes the ability to opt out of having your data used for model improvement.

What you can't control

Some aspects of data collection are unavoidable if you want to use the service, and being realistic about this helps you make informed choices.

You can't use AI without your messages going to company servers because that's how the technology works - the AI model is too large to run on your device, so everything has to be processed remotely. If you're not comfortable with that, you can't use cloud-based AI services.

You can't completely prevent metadata collection because some information (your IP address, device type, when you access the service) is inherent to how internet services function. Even if you use privacy tools like VPNs to mask your IP address, some metadata will still be collected.

You can't verify what companies actually do with your data beyond what they claim in their policies. You're trusting them to follow their stated practices, and while most large companies do (because violating their privacy policies creates legal liability, which could put them out of business or at least in serious financial trouble), you can't independently audit their data handling.

You can't prevent data breaches because even companies with strong security can be compromised. If you've used an AI service, your data could potentially be exposed, just like with any online service.

Practical steps to protect your privacy

Here's what you can actually do to limit how much information AI companies collect and store.

Never type sensitive personal information into AI conversations, including passwords, financial details, medical information, or anything you wouldn't want potentially exposed. Assume that anything you type could theoretically be seen by someone else, and make decisions accordingly.

Use throwaway email addresses for accounts if you want to use AI services but limit how much personal information is tied to your usage. Free email services make it easy to create separate accounts for different purposes. Personally, I have one generic account which if it was compomised would be no big deal and I use that email address most of the time. I have another one which is only shared with financial institutions, governments and the like. I have a few more to maintain and track project work independently. the only caveats are that you must NOT share passwords across accounts and ideally have proper account security and safe recovery methods for them all. But that's me - I'm a bit of a nerd in case you hadn't guessed.

Review and delete old conversations periodically because many services let you view and delete your conversation history. Make this a regular habit if you're using AI frequently for work or personal tasks, and it's also good housekeeping to clear up clutter. If there is 'stuff' you really might need one day (and we all have 'stuff' like that!) then consider copying into a Notes app or similar and password protect that,

Finally, check privacy settings in your account because most AI services have options to control data retention, opt out of using your data for training (if that option exists), or limit certain types of data collection.

The bigger picture

AI companies collect less data about you than social media platforms or advertising networks, but they do collect the conversations you have with them, and those conversations can contain a lot of personal information if you're not careful about what you share.

The question isn't whether AI companies collect data (they do!) but whether you're comfortable with what they collect and how they use it. For casual use like asking AI to explain concepts, draft emails, or answer general questions, the privacy implications are relatively minor.

For sensitive work involving confidential information, personal details, or anything you wouldn't want potentially exposed, you need to be much more careful about what you type and which services you use.

The goal isn't to avoid AI entirely out of privacy concerns but to understand what data is being collected so you can make informed decisions about when and how to use these tools. Sometimes the convenience is worth the trade-off, and sometimes it isn't, and that calculation will be different for different people and different situations.

Browse all topics → Index