How to Spot AI-Generated Videos and Images Online

The Challenge of AI-Generated Content

• 💡 The internet is increasingly filled with AI slop, making it difficult to distinguish real images and videos from synthetic ones.
• 🧠 AI models have advanced, hallucinating less and producing more realistic content, reducing obvious visual cues like mangled text or incorrect finger counts.
• 📱 A significant factor is mindless scrolling through social media feeds, where users don't critically examine every piece of content.
• ⚠️ AI content can be intentionally designed to evoke emotional responses, bypassing critical thinking and making users less likely to question its authenticity.

Strategies for Identifying AI Content

• 🏷️ Look for visible watermarks or platform-specific labels that indicate content has been generated or modified by AI, as implemented by platforms like TikTok, Meta, and YouTube.
• 🔍 Utilize reverse image searches and check video metadata for signals that can alert you to AI generation, such as specific tags within Sora or Gemini videos.
• 🤔 Employ critical thinking by questioning the plausibility of what you see, using a "physics test" to determine if actions or appearances defy natural laws (e.g., a cat jumping 50 ft).

Combating AI Slop

• 📣 The movement against AI slop involves collective effort, encouraging users to be more discerning and push for better detection tools.
• 📢 The clearest way to identify AI content is for creators to be transparent and label their AI-generated posts, with the implication that undisclosed AI use may be questionable.
• 🌐 For more in-depth information, explore resources like CNET's AI Atlas and their full story on fighting AI slop.