Five Tools to Verify If an Image or Article Was AI-Generated
Practical guides to protect yourself, your family, and your business from AI-driven scams, deepfakes, and emerging cyber threats.
Popular articles
The verification habit
AI-generated content is no longer a curiosity. It is the default for spam, propaganda, fake news, and a growing share of social media engagement bait. Spending twenty seconds to verify before you share is the single most valuable digital hygiene habit you can build.
Here are five free or low-cost tools that cover the main content types.
1. AI or Not (aiornot.com)
For images. Upload or paste a URL. Returns a probability score. Strong on photographs and portraits, weaker on illustrations and edited images. Free tier is generous.
2. Hive Moderation Demo (hivemoderation.com/ai-generated-content-detection)
For images, video, and audio. The same engine that runs commercial moderation pipelines at Reddit and other platforms. Browser demo is free for individual checks.
3. Originality.AI
For text. Paid (cents per check) but built for publishers and SEO teams. Detects ChatGPT, Claude, Gemini, and most open-source models with decent accuracy on longer text.
4. C2PA Verify (contentcredentials.org/verify)
For provenance, not detection. If a photo has C2PA Content Credentials attached — increasingly common for Adobe, Leica, Sony, and major news outlets — this tells you who created it and what edits were made. The future of trust online runs through provenance, not detection.
5. Reverse image search (Google Lens, TinEye)
The oldest tool is still one of the best. If a 'breaking news' photo has appeared in a 2018 article, you have your answer. Most viral misinformation is recycled or repurposed.
How to use them
No single tool is reliable on its own. The workflow that actually works:
- Reverse-image-search first. Confirms whether the image is real but old.
- Run a detector if the image is new or unfamiliar.
- Check provenance via C2PA if available.
- If two or more signals say 'likely AI,' do not share. Find a verified source instead.
What detection cannot do
Detectors are an arms race. They get better; generators get better; the gap closes. Treat any tool's verdict as a probability, not a proof. The real defense is a habit of skepticism toward anything emotionally provocative on social media. If a piece of content makes you want to share it immediately, that is the moment to slow down.