Unseeable prompt injections in screenshots: more vulnerabilities in Comet and other AI browsers
35
votes
-
-
Highlights from the Claude 4 system prompt
25 votes -
Criminals are getting increasingly adept at crafting malicious AI prompts to get data out of ChatGPT
22 votes -
Boffins (CMU) build automated method to bypass LLM guardrails
8 votes -
Prompt injection: What’s the worst that can happen?
8 votes -
The prompt box is a minefield
11 votes -
Prompt injection attacks against GPT-3
14 votes