The dense “terms and conditions” that most users scroll past when opening a social media account contain sweeping permissions to collect user data and limit users’ ability to take companies to court.Those are some of the conclusions from a new Harvard University research tool that documents old and current versions of legal documents.The so-called Transparency Hub stores over 20,000 documents and tracks the terms of over 300 platforms, including TikTok and Instagram.The goal of the platform is to make it easier for people to know where their data is going and what their rights are, according to Jonathan Zittrain, professor of international law at Harvard.  One finding is that these documents are getting harder to understand over the years. Using a readability metric called the Flesch-Kincaid Grade Level, researchers analyzed privacy policies from 2016 to 2025 and discovered that about 86 percent now require a college-level reading ability.The tool comes as European countries such as France, Portugal, Spain and Denmark decide what restrictions, if any, to put on social media to limit children’s harmful use.A shift away from the courtroomAnother emerging pattern is that many platforms are steering disputes away from public courts.Users are often required to settle conflicts through arbitration: a private process in which a neutral third party issues a binding decision, according to a statement from Kevin Wrenn, a researcher with Boston University, who used the Transparency Hub.Wrenn said in most cases, these companies also pick mediators themselves to settle them, which quietly removes the user’s right to sue them in court.Current terms and conditions for AI platforms such as Anthropic and Perplexity have notices that say that users cannot participate in a class action lawsuit against the company.This forces users who suffer damages from using the platforms to bring legal action individually, rather than pursuing claims collectively before a judge or jury. Users can opt-out of Perplexity’s restrictions by sending a written notice to a support email within 30 days of first using the AI, according to the terms of service.Euronews Next reached out to Anthropic and Perplexity for comment on their arbitration policies, but did not receive an immediate reply.It is not immediately clear whether the terms and conditions differ for European users compared with those in the United States.