Since August 2, 2026, the **AI Act** has mandated that certain AI-generated content be clearly labeled in the European Union. This regulation requires that deepfakes, informational texts produced without human editorial oversight, and chatbots be identified as such, with penalties that can reach up to 15 million euros or 3% of a company’s annual global turnover. To help verify the origin of such content, 12 free tools have been selected to analyze and report on AI-generated text, images, or videos. Several major platforms have introduced features to label AI-generated content. Instagram renamed its "AI Creator" label to "AI-Generated Profile" on August 31, which applies only to accounts whose main character is an AI-generated person. WhatsApp now allows channel administrators to label AI-generated content, which appears at the bottom left of a post and cannot be removed once confirmed. This feature is currently available only to a portion of beta testers on Android and iOS. Adobe has introduced a free tool called Content Authenticity, still in beta, which allows users to add Content Credentials to images, audio files, or videos. These metadata, following the open standard C2PA, provide information such as the creator’s name, their social media accounts, the tools used, and whether generative AI was involved in the creation. This helps users trace the origin of digital content more transparently. LinkedIn and YouTube have also provided mechanisms for users to flag AI-generated content. LinkedIn introduced the "Seems like AI slop" button, accessible via the three dots at the top right of a post or comment. Reports from this button help improve the platform’s detection models. YouTube’s detection of similarity feature, available to users over 18, compares a biometric signature of a video with previously uploaded content and alerts users if a match is found. This allows users to request the removal of potentially AI-generated content. Numerous tools analyze content to detect AI-generated material. GPTZero, launched on January 2, 2023, allows users to check up to 10,000 words per month in its free plan. Copyleaks checks for both AI-generated content and plagiarism, covering over 30 languages for AI content and more than 100 for plagiarism. Pangram estimates whether a text or image was created by a human, an AI, or with AI assistance. Hive AI Detector is a free Chrome extension that analyzes content directly from the browser, while AI or Not focuses on image analysis, detecting AI-generated images regardless of their origin. Companies like Anthropic and OpenAI have developed their own verification tools. Anthropic launched a public verification tool called Check if a file was made with Claude, which reads files produced by their models. OpenAI uses a combination of C2PA metadata and an invisible watermark called SynthID to verify images and audio files. In October 2026, OpenAI extended this marking to text with textGrain, which adds an invisible statistical signal to the model’s word choices. Despite these tools, none offer absolute certainty, and their developers do not claim they can definitively establish the origin of content. These tools are trained on existing data and thus lag behind each new AI model release. A study from June 2026 by researchers from Cornell Tech and MIT found that detection performance dropped in the weeks after the release of GPT-5.4. OpenAI also admits that rewriting a text can weaken its watermark. These tools are useful for identifying content that may need further scrutiny, but they cannot conclusively determine its origin.