Presentation Material
Abstract
This talk is inspired by an episode of Black Mirror. I will be demonstrating a live demo creating a bot who talks like me and can be used to impersonate me online and do social engineering. I will be showing a live demo of how to a create such bots over text, voice or video and walk through various techniques which the attendees can use to create such smart social engineering attacks.
I will also release my github of the AI notebooks as open source for the attendees to try out and experiment.
AI Generated Summarymay contain errors
Here is a summarized version of the content:
The speaker presents a prototype of an AI-powered chatbot that uses lip syncing, , speech-to-text recognition, and video streaming to create a realistic conversation experience. The system consists of several components:
- Brain: generates responses
- Voice Engine: converts text to voice
- Face Engine: imposes voice on a face video with lip syncing
- Video Streamer (Cam Twist): streams silent video and switches to talking video when responding
The speaker demonstrates the prototype in a call between their bot and another person, showing how it can have a conversation.
To improve the system, the speaker plans to:
- Make the voice more convincing
- Add facial expressions and emotions
- Bypass biometric security measures
The speaker is also working on a course or workshop to teach others about this technology. They invite support from angels and other interested parties to help build this futuristic tech.
Finally, they thank the audience for their time and invite further questions via email or LinkedIn.