AI or Not

AI or Not helps you tell human-made content from material generated by AI models. In one interface, you can check text, images, audio, video, and potential deepfakes—useful for newsrooms, brands, educators, and individual users.

What you can scan

• Text pasted into the tool
• Uploaded images and graphics (PNG, JPG, WEBP, GIF)
• Video files (MP4 and other common formats)
• Audio files (MP3 and other common formats)

How it works

Upload a file or paste text, and the system analyzes it for signals of AI generation. The service uses specialized models trained on content associated with tools such as ChatGPT, Claude, Midjourney, Suno, Udio, Veo, Kling, and similar systems.

Deepfake detection and brand safety use cases

• Flag suspected deepfake video and audio
• Check image authenticity
• Monitor potential unauthorized brand use
• Support moderation of user-generated content
• Review student work and written assignments
• Audit marketing and ad materials
• Personally verify suspicious files

The developers report high detection accuracy based on open academic datasets containing real and synthetic images.