🔬 Science 4d ago · Frank Landymore

Simple Prompt Turns ChatGPT Into a Sociopath That Ignores Safety Guardrails

Futurism
Futurism science news
View Channel →
Simple Prompt Turns ChatGPT Into a Sociopath That Ignores Safety Guardrails
Source ↗ 👁 0 💬 0
Researchers at the British AI security startup Mindgard found that a simple prompt spurred ChatGPT to drop its most basic safety guidelines, in another example of how the guardrails surrounding even the most popular AI models can easily be circumvented.



Specifically, according to reporting from the BBC, they coaxed OpenAI’s model to generate gruesome photorealistic scenes depicting gore and sexual content. Mindgard’s technique only involved slightly changing a widely-shared prompt

Comments (0)

Sign in to join the discussion

More Like This

Restored WWII fighter plane tests out .50 caliber machine guns
Popular Science · 12h ago
Mysterious Compound Detected on Pluto and Titan
Science Latest · 13h ago
Ancient rocks suggest water has shaped earth for 3.1 billion years
Popular Science · 14h ago
Super rare orange lobster molts at New York Aquarium
Popular Science · 15h ago
Dragonflies maneuver like fighter pilots
Science - Ars Technica · 15h ago
AI labels a lot of stuff as alien life
Popular Science · 16h ago