It's Already Happening
LinkedIn, Meta, X, and dozens of other platforms are feeding your posts, photos, and data into AI training systems. Most turned this on by default. Most users never knew.
If you didn't opt out, your data is already in the machine. And once it's trained into a model, there's no getting it back.
The Quiet Data Grab
Here's what's happening right now across major platforms:
LinkedIn (Microsoft)
Status: Training on your data by default
Deadline: November 3, 2025 for EU/EEA/UK users [1]
US users: Already being scraped
X (Twitter/Grok)
Status: Training Grok on all public posts
Third parties: Now allowed to train on your data too [2]
Mobile opt-out: Not available
Meta (Facebook/Instagram)
Status: Training on public posts
EU users: Started May 27, 2025 [3]
US users: No opt-out option
Adobe
Status: Changed terms in 2024
Content: Can access your Creative Cloud files
Training: "Content analysis" for AI
LinkedIn: The Microsoft AI Pipeline
LinkedIn announced on September 18, 2025 that it would share user data with Microsoft and its affiliates for AI training. [1]
This isn't asking for permission. It's "legitimate interest"—corporate speak for "we're doing it unless you stop us."
What They're Taking
Everything public on your profile: [4]
- Your name, photo, and bio
- Current and past work experience
- Education history
- Skills and endorsements
- Publications and patents
- Recommendations you've received
- Every post and comment you've made
- Job application data (resumes, screening responses)
Private messages are excluded. Everything else is fair game.
The Deadline Problem
LinkedIn gave EU, EEA, UK, Switzerland, Canada, and Hong Kong users until November 3, 2025 to opt out. [5]
For US users: your data has already been scraped for "some time." [1]
And here's the catch: even if you opt out after the deadline, the data collected before opting out stays in their training datasets permanently. [4]
How to Opt Out
- Go to Settings & Privacy
- Select Data Privacy
- Find "Data for Generative AI improvement"
- Toggle off "Use my data for training content creation AI models"
While you're there, also disable: [4]
- "Ads off LinkedIn"
- "Data from others for ads"
- "Measure ad success"
- "Share data with affiliates and partners"
X (Twitter): Feeding Grok and Third Parties
Elon Musk's X has been training Grok on user posts since 2024—and doing it quietly.
The Silent Toggle
In July 2024, X activated a setting that, by default, gives it access to users' posts, interactions, and inputs as training data for Grok. [6]
Most users never knew. The setting was buried in privacy options and enabled without notification.
What They're Taking
- All your public posts
- Likes and interactions
- Spaces participation
- Profile information (bio, display name)
- Your Grok conversations
Private accounts are exempt. But if you've ever posted publicly, that content is now Grok training data.
Third Parties Get Access Too
In October 2024, X updated its privacy policy to allow "third-party collaborators" to train their AI models on X data. [7]
Not just Grok. Any company X partners with can now use your posts for AI training.
The Mobile Lockout
You can't change this setting in the mobile app. Desktop only. [6]
To opt out:
- Log into X on desktop (not mobile)
- Go to Settings → Privacy and Safety
- Select "Data Sharing and Personalization"
- Click "Grok"
- Toggle off
Musk's Goal
xAI is aiming to make Grok the world's top AI by December 2025, using what Musk calls "the most powerful AI training cluster." [8]
Your tweets are the fuel.
Meta: Facebook and Instagram
Meta delayed its European AI training plans after regulatory pushback, but as of May 27, 2025, it started using EU users' public Facebook and Instagram data for training. [3]
US Users Have No Choice
For users outside the EU, there are no data privacy laws like GDPR. Meta has been training on your public posts with no opt-out option. [3]
Your photos. Your comments. Your life updates. All feeding Meta's AI models.
The Consent Theater
EU users got an "opt-out" notice. US users got nothing. Same company. Same data. Different rules based on where you live.
The Pattern Is Clear
Every major platform is doing the same thing:
- Enable AI training by default
- Bury the opt-out in settings
- Give minimal notice
- Make it irreversible once collected
They're not asking permission. They're counting on you not noticing.
Why This Matters
Your Data Becomes Their Product
AI models trained on your content are worth billions. Microsoft, Meta, and xAI are building products they'll sell to enterprises—using your words, your expertise, your creativity as raw material.
You don't get paid. You don't even get credit.
You Can't Take It Back
Once data is trained into a model, it's there permanently. There's no "unlearning" process. Deleting your posts after the fact doesn't remove them from the model weights.
Every professional insight you shared on LinkedIn. Every hot take on X. Every family photo on Instagram. Baked into AI forever.
Privacy Violations Compound
AI models can be interrogated to extract information. Researchers have demonstrated "inference attacks" that pull out data about the training set. [9]
Your data doesn't just train the model. It can potentially be extracted from it.
The Two-Tier System
Notice who gets protection and who doesn't:
- EU users: Got deadlines, notifications, and opt-out mechanisms (GDPR)
- US users: Got scraped without warning and have no opt-out
Same platforms. Same data. Different rights.
LinkedIn explicitly states that US users' data has "already been scraped for some time." [1] Meta doesn't even pretend to offer US users a choice. [3]
What You Can Do
- Settings & Privacy → Data Privacy
- Toggle off "Data for Generative AI improvement"
- Also disable data sharing with affiliates
- Do this now — data collected before opt-out stays in training sets
X (Twitter)
- Log in on desktop (not mobile)
- Settings → Privacy and Safety
- Data Sharing and Personalization → Grok → Off
Meta (Facebook/Instagram)
- EU users: Look for AI data usage settings in Privacy Center
- US users: No opt-out available. Consider what you post publicly.
General Protection
- Make profiles private where possible
- Limit public posting — especially professional expertise on LinkedIn
- Delete old content before opt-out deadlines (may not help once scraped)
- Assume everything public will be used
The Bigger Picture
This is the new business model for social media:
- Build a platform where people share their lives
- Harvest that content to train AI
- Sell AI products to enterprises
- Give users nothing but a terms of service update
You're not the customer. You're not even the product anymore. You're the raw material.
The Uncomfortable Truth
You Built Their AI For Free
Every LinkedIn post demonstrating your professional expertise? Training data for Microsoft's Copilot.
Every tweet sharing your opinions? Training data for Grok.
Every Instagram photo of your life? Training data for Meta AI.
These companies are building billion-dollar AI products using your content as the foundation. They gave themselves permission through terms of service updates you probably didn't read.
The opt-outs exist, but they're designed to be missed. And for US users, some platforms don't even pretend to offer a choice.
Welcome to the data economy. Your words have value—just not to you.
References
- Malwarebytes - LinkedIn will use your data to train its AI unless you opt out now
- TechCrunch - Elon Musk's X is changing its privacy policy to allow third parties to train AI on your posts
- Proton - LinkedIn will use your data to train AI – how to opt out
- Tuta - Act until Nov 3rd to stop LinkedIn from abusing your data to train AI!
- The Register - One week to opt out or be fodder for LinkedIn AI training
- Variety - Elon Musk's X Activates Setting Allowing It to Train Grok AI on User Posts by Default
- Euronews - Elon Musk's X quietly changes default settings to train Grok with your posts
- AppleInsider - How to stop Musk's Grok AI from training on anything you've said on X
- Stanford Report - Study exposes privacy risks of AI chatbot conversations
- LinkedIn Help - Update to our Terms and data use