91% of people accept Terms of Service without reading them. Reading every TOS you encounter in a year would take 76 full work days. Microsoft's alone is 15,260 words. Companies know this. They count on it. And when nobody reads the fine print, extraordinary things happen. The Soul Clause (2010) In April 2010, UK retailer GameStation added a clause to their terms: by placing an order, you granted them "immortal soul" rights. It was an April Fool's joke, but 88% of customers agreed without noticing. The point was made. Nobody reads this stuff. Apple Will Not Let You Build Nukes Buried in Apple's iTunes EULA, Section G explicitly forbids users from using the software "for the design, development, fabrication, or testing of nuclear, chemical, or biological weapons." Nobody at Apple genuinely worried about nuclear proliferation via iTunes. But the clause exists because lawyers put it there, and millions of people agreed to it without reading a single word. Instagram Tries to Sell Your Photos (2012) In December 2012, Instagram updated its terms to claim the right to sell user photos to advertisers without compensation or notification. The backlash was immediate and massive. Within 48 hours, Instagram backtracked. But the episode revealed something important: the legal framework to monetize your content was already built. They just moved too fast. WhatsApp Forces Data Sharing With Facebook (2021) In January 2021, WhatsApp told its 2 billion users: share your data with Facebook, or lose access to your account. The "choice" was between consenting to data harvesting or being cut off from your contacts. Millions migrated to Signal and Telegram. But most users simply clicked "Agree." Zoom Claims Your Content for AI (2023) In August 2023, Zoom updated its terms to claim a perpetual, irrevocable license to use customer content — including video calls — for AI training. After public outrage, Zoom clarified that "consent-based" meant something different than what the plain text said. If you had a confidential meeting on Zoom that month, your content was covered by terms that said Zoom could use it. Adobe Enters Your Creative Cloud (2024) In June 2024, Adobe updated its Creative Cloud terms to require access to customer content. Photographers, designers, and artists who stored work in the cloud were told that Adobe needed to review their files for "content moderation." Professionals with confidential client work in Creative Cloud had already agreed. Disney+ Blocks a Wrongful Death Lawsuit (2024) In August 2024, Disney tried to use the Disney+ terms of service to block a wrongful death lawsuit. A man's wife died from anaphylactic shock at a Disney restaurant. Disney argued that because he had signed up for a Disney+ free trial years earlier, he had agreed to arbitrate all disputes. Disney later backed down. But the legal argument was real, and it nearly worked. LinkedIn Opts Everyone Into AI Training (2024) In September 2024, LinkedIn quietly updated its settings to opt all users into AI training using their posts, articles, and messages. There was no announcement. No email. Just a settings toggle that was flipped to "on" by default. Users had to manually navigate to Settings > Data Privacy > Data for Generative AI Improvement and turn it off. Most never did. TikTok's Forced Terms (2026) In January 2026, TikTok forced a new terms of service on users. The twist: you couldn't delete your account without first accepting the new terms. Users were locked into an agreement they wanted to reject just to exercise their right to leave. TikTok also collects keystroke patterns — the rhythm and timing of how you type — as "automatically collected information." X/Twitter Sells Your Posts for AI Training (2024) In November 2024, X's updated terms allowed third-party AI companies to train on user tweets. Every post, reply, and DM became potential training data for AI models you've never heard of, built by companies you'll never know. Meta's Automatic AI Consent (2025) In December 2025, Meta updated its privacy policy to confirm that all Facebook and Instagram posts are used for AI training. There is no opt-out for users in most regions. Your photos, your writing, your conversations — all grist for the AI mill. Anthropic's Bait and Switch (2025) In August 2025, Anthropic — which had built its brand on "we don't train on your data" — shifted to an opt-out model with a 5-year data retention policy. Users who trusted Anthropic's original commitment found that the terms had changed underneath them. The company that sold itself as the ethical alternative quietly adopted the same data practices it once criticized. Reddit Sells Your Data for $60M/Year Reddit sells user-generated content to Google for AI training at approximately $60 million per year. The users who wrote the posts, built the communities, and created the value see none of that revenue. They agreed to the terms, though. Somewhere in the signup flow, there was a link they clicked. The Numbers 91% of consumers accept TOS without reading (Deloitte) 76 work days per year would be needed to read every TOS 15,260 words in Microsoft's terms alone The average TOS requires college-level reading comprehension What You Can Do Use ToS;DR — Terms of Service; Didn't Read rates services and highlights problematic clauses Check before you sign up — Search "[service name] terms of service controversy" before creating accounts Use privacy-first alternatives — Services that don't train on your data exist for most categories Delete unused accounts — Every dormant account is still governed by terms you agreed to Support regulation — GDPR-style consent requirements and plain language laws make terms readable and enforceable The Real Cost Every "I Agree" button is a contract. Every click binds you to terms written by teams of lawyers to protect companies, not you. The system is designed so that reading is impractical and declining is often impossible. They didn't ask if you understood. They didn't ask if the terms were fair. They just needed your click. --- Sources: GameStation/PCGamer 2010, Apple EULA Section G, Instagram/New York Times Dec 2012, WhatsApp/TechCrunch Jan 2021, Zoom/THE VERGE Aug 2023, Adobe/The Washington Post June 2024, Disney+/CNBC Aug 2024, LinkedIn/404 Media Sept 2024, TikTok/Mashable Jan 2026, X/THE VERGE Nov 2024, Meta/Privacy Policy Dec 2025, Anthropic/Ars Technica Aug 2025, Reddit/DealBook $60M Google deal, Deloitte Digital Democracy Survey.