03.10.2024 -TikTok, Instagram, Snapchat and other similar apps offer many features that use artificial intelligence (AI), such as image generators, chatbots and tools that make AI-generated human faces. When an image or video is uploaded to the app, it is automatically analysed to suggest effects, filters or subtitles. These AI-driven background functions are not problematic in themselves, but their risks become apparent when they are used by criminals to manipulate their victims.
AI has been used in social media for some time to optimise user experience. It is used, for example, to improve images. Young people today use AI features as a matter of course, often without giving it much thought.
AI image generators let you modify or create photos, videos and audio files from scratch. Deepfakes are the result. Currently, it is usually possible to identify a deepfake on closer inspection.
Typical signs of audio deepfakes (also called voice cloning) are:
- The person's voice sounds metallic and monotonous
- The content of the conversation seems strange
- The person is uses words they don't normally use
Typical signs of video deepfakes are:
- Strange shadows or a haircut that does not match the face
- Transitions between parts of the face appear blurry
- Differences in resolution (person vs. background)
Cybercriminals use deepfake technologies to disguise their identity and establish a level of trust with the victim. Children and young people are vulnerable to this kind of manipulation because many of their social interactions take place online. Young people are often more trusting than adults. In the video below, Brian Ceccato, technical analyst at the National Cyber Security Centre, shows how numbers can be spoofed, voices and even Swiss dialects can be imitated, and live videos can be faked.
Tips:
- Protect yourself and your data
- Take the time to look closely
- Be suspicious, even online
- Check the sources
- Be cautious when you publish personal information
Further Information
Last modification 03.10.2024
