all AI news
Indirect Prompt Injection via YouTube Transcripts
May 15, 2023, 7:11 p.m. |
Simon Willison's Weblog simonwillison.net
Indirect Prompt Injection via YouTube Transcripts
The first example I've seen in the wild of a prompt injection attack against a ChatGPT plugin - in this case, asking the VoxScript plugin to summarize the YouTube video with ID OBOYqiG3dAc is vulnerable to a prompt injection attack deliberately tagged onto the end of that video's transcript.
Via danShumway
ai case chatgpt example generativeai llms plugin prompt prompt injection promptinjection security transcripts video vulnerable youtube
More from simonwillison.net / Simon Willison's Weblog
Merge pull request #1757 from simonw/heic-heif
1 day, 1 hour ago |
simonwillison.net
Wrap text at specified width
1 day, 3 hours ago |
simonwillison.net
llm cmd undo last git commit - a new plugin for LLM
2 days, 15 hours ago |
simonwillison.net
GGML GGUF File Format Vulnerabilities
2 days, 23 hours ago |
simonwillison.net
Jobs in AI, ML, Big Data
Senior ML Researcher - 3D Geometry Processing | 3D Shape Generation | 3D Mesh Data
@ Promaton | Europe
Principal Data Engineer
@ RS21 | Remote
SQL/Power BI Developer
@ ICF | Virginia Remote Office (VA99)
Senior Machine Learning Engineer (Canada Remote)
@ Fullscript | Ottawa, ON
Software Engineer - MLOps.
@ Renesas Electronics | Toyosu, Japan
Junior Data Scientist / Artificial Intelligence consultant
@ Deloitte | Luxembourg, LU