What makes media a deepfake?
A deepfake uses AI to generate or alter a photo, video, or audio recording. It can make a person appear to say or do something that never happened, or show their likeness in an invented scene. Synthetic media is a broader term that also covers original AI-generated content, such as a portrait of a fictional character.
Permission and context matter when a real person appears in altered media. For example, get a participant’s consent for a film effect and label the altered clip where viewers could take it as a real recording. A fabricated product endorsement or payment request can mislead viewers and harm the person depicted.
How to check a suspicious image, video, or recording
Look for the original post or recording and check whether the person or organization has confirmed it through a channel you already trust. If a clip or voice message asks for payment, account access, or personal details, contact the sender through a known number or address before acting.
Visual defects can come from editing or ordinary compression. Treat detection results as leads and verify the source independently. When available, content credentials can show the recorded origin and edit history. Authentic media can also circulate without those records.
Real-person likenesses and fictional characters on Dreamgens
Dreamgens prohibits real-person sexual deepfakes, sexual impersonation, and non-consensual intimate depictions. Use Report Abuse for a suspected policy violation on Dreamgens, and include the relevant page link so the team can locate it.
For a series of original character images, start with a fictional adult character you created or have permission to use. Character Reference uses views of that character to guide a new scene, and the reference images and outputs stay private. When sharing work elsewhere, identify it as AI-generated where viewers could mistake it for a real event or person.