As major tech companies accelerate the rollout of artificial intelligence tools across their platforms, many users in the United States are increasingly anxious about how much of their personal data is being accessed, analysed, or used to train these systems. With companies like Meta, Google, and LinkedIn introducing new AI-driven features, questions continue to rise about privacy protections, transparency, and the ability to opt out.

In recent weeks, viral posts on Instagram, X (formerly Twitter), and other platforms have sparked concerns that tech companies are now harvesting private messages, emails, photos, and voice recordings to feed their AI models. But how much of this is true, and what are tech companies actually doing with your information?

Below is a breakdown of the facts.

---

Why Users Are Worried About AI and Privacy

Tech companies are not always transparent about the types of data they collect or how they use it, said Krystyna Sikora of the Alliance for Securing Democracy. This lack of clarity allows misinformation to spread quickly and fuels public fears about AI systems reading private conversations or pulling sensitive content without consent.

The United States lacks a comprehensive federal privacy law, creating wide gaps in user protections compared with regions like the European Union or the United Kingdom.

To understand what’s really happening, it’s important to look at each platform’s policies.

---

Meta: Can Facebook, Instagram, and WhatsApp Use Your Data for AI?

Viral Claim

“Starting December 16, Meta will read your DMs, every photo, every voice message to train AI.”

The Facts

Meta is updating its policy on December 16, but this does not mean the company is scanning all private messages to train AI.

Here’s what Meta is doing:

• Meta’s AI tools do not use private messages from Instagram, WhatsApp, or Messenger for training.
• Meta does use public content (public posts, comments, reels, profile photos, etc.) to train AI models.
• Conversations with Meta’s AI chatbot may influence what content or ads users see.
• Meta AI may learn from information that appears in public posts—even if the person mentioned does not have a Meta account.

Can you opt out?
No. Meta does not offer a global opt-out for AI training.
You can disable the Meta AI feature in WhatsApp, but only for individual chats.

Facebook and Instagram users cannot turn Meta AI off entirely.

Meta confirms that deleting your account does not remove previously public data from being used in AI training.

---

Google: Is Gmail Allowing AI to Read Your Emails?

Viral Claim

“Google AI can now read every Gmail message and attachment.”

The Facts

Google’s new AI tool, Gemini Deep Research, can access your Gmail, Drive, and Chat content—but only if you explicitly give permission.

Google collects data in several ways:

• Through Gemini when users allow access to Gmail, Drive, or Chat
• Through interactions in Gemini apps
• Through uploads (images, videos) users provide to Gemini
• Through YouTube and other app activity (if permissions are granted)
• Through smart features in Gmail and Workspace (which are turned on by default in the US)

A recent lawsuit in California argues that Google changed its policies in October to give Gemini broader default access, requiring users to disable the settings manually. Google denies improper data collection.

Can you opt out?
Partially.

• You can turn off smart features in Gmail to stop AI from scanning email content.
• You can use temporary chats or sign out to prevent Gemini from saving your prompts.
• You cannot prevent Google from training AI on data collected from other Google services you actively use unless you disable specific permissions.

Google says it does not use data from users under age 13 to train AI.

---

LinkedIn: Will Your Profile Be Used to Train Microsoft’s AI Models?

Viral Claim

“LinkedIn will start using your data to train AI on November 3.”

The Facts

LinkedIn confirmed that it is using some user data to train generative AI models, including:

• Public posts
• Profile information
• Content users share publicly

LinkedIn does not use private messages for AI training.

Additionally, Microsoft now receives certain LinkedIn data—such as ad engagement and newsfeed activity—to personalise advertising.

Can you opt out?
Yes.
LinkedIn offers clear opt-out settings for:

• AI training
• Personalised advertisements
• Data sharing with Microsoft and partner companies

Users can find these options under Data Privacy → Data for Generative AI Improvement.

---

What Users Can Do to Protect Their Privacy

1. Review your privacy settings

Every major platform offers settings to limit data access—even though they’re often hard to find.

2. Disable features that allow AI access

For example, turning off Gmail’s smart features significantly reduces AI’s reach into your email content.

3. Be cautious with public posts

Public content—even tags or photos uploaded by other people—can be used to train AI models.

4. Understand app permissions

Many apps request microphone, photo, contacts, or location access. If you don’t need the feature, revoke the permission.

5. Consider using alternatives

Some platforms, including privacy-focused email services or messaging apps, do not use data for AI training at all.