What are the most common types of manipulated media encountered on Twitter?

On Twitter, manipulated media can take various forms, each designed to deceive, mislead, or influence public opinion. Here are some of the most common types of manipulated media encountered on the platform:

### **1. **Deepfakes**

- **Definition**: Deepfakes are videos or audio recordings that use artificial intelligence (AI) to create realistic but fake content. They can superimpose a person's face onto someone else's body or mimic someone's voice with high accuracy.
- **Impact**: These can be used to create misleading content or fake endorsements, potentially leading to misinformation and reputational damage.

### **2. **Image Manipulation**

- **Photo Editing**: Images may be altered using software to change or remove elements, create misleading impressions, or falsely depict events. Common manipulations include adding or removing people, changing backgrounds, or altering facial expressions.
- **Meme Generation**: Edited images are often turned into memes to spread misinformation or satire, sometimes distorting facts to fit a specific narrative.

### **3. **Video Manipulation**

- **Selective Editing**: Videos may be edited to cut out context, misrepresent statements, or create misleading impressions of events. This can include splicing together different clips to alter the original meaning.
- **Misleading Captions**: Videos can be posted with misleading or false captions that distort the actual content of the video.

### **4. **Textual Manipulation**

- **Fabricated Screenshots**: Screenshots of tweets, messages, or documents may be edited to misrepresent statements or create false evidence. These can be used to falsely attribute statements or create misleading narratives.
- **Fake Accounts and Bots**: Automated accounts may generate and spread text-based misinformation, including fake news stories or altered quotes.

### **5. **Audio Manipulation**

- **Voice Cloning**: AI tools can create realistic audio recordings that mimic someone's voice. These can be used to falsely attribute statements or create fake interviews.
- **Audio Editing**: Audio clips may be edited to remove context, alter statements, or create misleading impressions of what was said.

### **6. **Synthetic Media**

- **AI-Generated Content**: Content created entirely by AI, such as generated images or videos that depict fictional events or people, can be used to spread misinformation or create misleading narratives.
- **Synthetic Voices**: AI-generated voices can be used to create fake recordings or manipulate existing audio.

### **7. **Contextual Manipulation**

- **Out-of-Context Sharing**: Content is taken out of its original context and shared in a way that distorts its intended meaning. This can involve sharing old content as if it were recent or presenting content from one topic as related to another.
- **Misleading Framing**: The framing of content can be manipulated to mislead viewers, such as presenting biased or selective viewpoints without providing full context.

### **8. **Hoaxes and Fake News**

- **False Narratives**: Entirely fabricated news stories or hoaxes can be spread through manipulated media, creating false narratives or inflaming public opinion on various issues.
- **Satirical Content**: Satirical or parody content, while not always intended to mislead, can sometimes be mistaken for factual information, leading to confusion and misinformation.

### **9. **Impersonation and Fake Accounts**

- **Impersonation**: Fake accounts or impersonation of public figures can spread manipulated media or misinformation under the guise of authenticity.
- **Bot Activity**: Automated accounts (bots) can amplify manipulated media, spreading it rapidly and giving it the appearance of widespread support or credibility.

### **10. **Altered Metadata**

- **Misleading Metadata**: Changing metadata (such as timestamps or location data) on images or videos can distort the perceived context of the media, leading to false conclusions about when or where it was captured.

By understanding these common types of manipulated media, Twitter aims to develop better tools and strategies for detection and moderation, ensuring a more accurate and trustworthy information environment for its users.