Source |
TroyHunt |
Identifiant |
6936171 |
Date de publication |
2022-09-16 17:38:52 (vue: 2022-09-16 20:10:02) |
Titre |
Twitter pranksters derail GPT-3 bot with newly discovered “prompt injection” hack |
Texte |
By telling AI bot to ignore its previous instructions, vulnerabilities emerge. |
Envoyé |
Oui |
Condensat |
bot derail discovered emerge gpt hack ignore injection” instructions its newly pranksters previous telling twitter vulnerabilities “prompt |
Tags |
Hack
|
Stories |
|
Notes |
|
Move |
|