https://t.co/dAH1a0bJLN I Tried Pranking People with Free Wi-Fi in New Zealand Overseas Y Papa Radio: Decoding the World #AISummary The Hidden Trap in Free Wi-Fi Terms of Service and How to Protect Your Personal Information 🔳 All six people agreed to adopt a cat via free Wi-Fi At a shopping mall in Auckland, New Zealand, free Wi-Fi was offered for about three hours with a hidden clause in the terms of service requiring users to agree to adopt a cat. All six users who connected clicked “Agree and Connect.” 🔳 Built the setup using a small computer A low-cost single-board computer, the Raspberry Pi, was used. It leveraged a “captive portal” system that automatically displays terms of service upon Wi-Fi connection, creating a simple interface that anyone could easily agree to. 🔳 Hidden clause in the terms: transfer ownership of a cat Inserted into Article 7, Section 3 of the nine-paragraph English-language terms of service was a provision requiring users to transfer ownership of their pet cat to the provider. For those without cats, the clause extended to any future cat, cat toys, feeding bowls, and even napping spots. 🔳 Average time to agree: 53 seconds The time taken by the six users to click “Agree” ranged from 38 to 61 seconds, averaging 53 seconds. The terms contained approximately 425 words—requiring about two minutes to read at a normal reading speed. All users agreed before finishing the text, though it’s unknown how much they actually read. 🔳 Experimental design collected no personal data The test Wi-Fi had no internet access and did not collect names, email addresses, or any personal information. Only connection and agreement timestamps were recorded. After agreeing, users were shown a disclosure screen raising awareness about the dangers of public Wi-Fi. 🔳 Similar experiment conducted in London in 2014 Finnish cybersecurity firm F-Secure inserted a clause into London’s free Wi-Fi terms stating users would permanently surrender their firstborn child. Six users agreed—this cat experiment was a replication of that earlier test. 🔳 2016 university study found 98% overlooked clauses Researchers from universities in the U.S. and Canada inserted clauses into a fictional social media platform’s terms of service requiring users to surrender their children or provide personal data. 98% of students missed these clauses; most spent less than one minute reading the terms. stringLiteral The structural problem: no one reads terms of service Carnegie Mellon University estimates that reading all privacy policies for websites used daily would take about 240 hours per year. The speaker argues the issue isn’t user negligence, but rather the systemic design that routinely demands consent without reading. stringLiteral The real danger lies in legitimate terms of service While cat or child surrender clauses are jokes, actual terms often contain hidden provisions for third-party data sharing, automatic renewals, and restrictive cancellation policies. By agreeing without reviewing them, users may unknowingly consent to data disclosures. stringLiteral Assume your personal information is already compromised The speaker urges people to act as if their name, email, phone number, and other details are already known to others. Just because someone has your personal information doesn’t mean they’re trustworthy. stringLiteral Prioritize trustworthiness of the provider over reading terms Since reading every term is impractical, choosing reputable companies and services is essential. Avoid easily agreeing to free Wi-Fi or services from unknown providers. stringLiteral Use AI to detect dangerous clauses The speaker demonstrates how AI can analyze lengthy terms of service to flag unusual clauses related to pricing, data handling, or cancellation policies. When this experiment’s terms were analyzed by AI, it detected the cat adoption clause and raised concerns about the provider’s credibility. stringLiteral The greatest lesson: the danger of the “Agree” button Hidden unfavorable conditions buried within long terms—and forcing consent to proceed—are commonplace. The speaker concludes that defending against data leaks requires: awareness of potential exposure, verifying provider trustworthiness, and using AI to review terms.