Alerts
Events
DCR
Explore Cyware Products
Alerts
Events
DCR
Go to listing page
Researcher Outsmarts, Jailbreaks OpenAI's New o3-mini
Malware and Vulnerabilities
February 08, 2025
Silent Push
Despite its improvements, a CyberArk researcher found a way to exploit o3-mini by pretending to be a historian seeking knowledge. While engaging with it, he eventually led it to produce steps that could be used to exploit a critical Windows process.
Read More
OpenAI
OpenAI o3-mini
Large Language Model (LLM)
Deliberative Alignment
Local Security Authority Subsystem Service (LSASS)
Publisher
Previous
DDoS Attacks Reportedly Behind DayZ and Arma Network Ou ...
Breaches and Incidents
Next
Critical RCE Flaw in Microsoft Outlook Now Exploited in ...
Malware and Vulnerabilities