Researchers demonstrate ChatGPT can be tricked into generating violent and sexual images
By
Chris Vallance
Summary
British AI security startup Mindgard discovered that the latest public version of ChatGPT can be tricked into generating sexualized and violent images by slightly altering a widely-shared prompt originally designed for humorous results. The researchers demonstrated that despite OpenAI's safety measures, simple prompt modifications can bypass content filters and produce graphic content including sex crime scenes. OpenAI acknowledged the issue and stated they are working to address the vulnerability.
Source
bskyResearchers demonstrate ChatGPT can be tricked into generating violent and sexual imagesbbc.comKey quotes
· 3 pulledThe latest public version of ChatGPT can be made to generate sexualised images or depict scenes of graphic violence with a simple prompt
British AI security startup Mindgard figured out how to make ChatGPT create graphic pictures by slightly altering a widely-shared instruction
Researchers say it is still possible to trick the AI chatbot into producing graphic content
You might also wanna read
OpenAI indefinitely shelves plans for ChatGPT 'adult mode' amid concerns about harmful effects
OpenAI has indefinitely shelved plans to release a sexualized 'adult mode' for ChatGPT after facing pushback from employees and investors co
The Verge·3mo ago
Study Finds AI Chatbots Vulnerable to Jailbreak Attacks Using Poetic Prompts
Researchers discovered that AI chatbots like ChatGPT can be tricked into providing dangerous information about nuclear weapons, child sex ab
wired.com·7mo ago
OpenAI Scans ChatGPT Conversations and Reports Threats to Law Enforcement
OpenAI has implemented a policy of scanning users' ChatGPT conversations and reporting threatening content to law enforcement. This comes am
futurism.com·10mo ago
OpenAI's Safety vs. Growth Dilemma: Balancing ChatGPT's Appeal with User Protection
OpenAI faced a dilemma between making ChatGPT more appealing to users and maintaining safety standards. The company initially tweaked its ch
nytimes.com·7mo ago
OpenAI Launches ChatGPT Images 2.0 with Web Search Capabilities for Enhanced Image Generation
OpenAI has launched ChatGPT Images 2.0, an updated version of its AI-powered image generator with new 'thinking capabilities' that allow it
The Verge·2mo ago
OpenAI Introduces Lockdown Mode to Protect ChatGPT Users from Prompt Injection Attacks
OpenAI is introducing Lockdown Mode for ChatGPT, a security feature that limits access to web browsing and external services to reduce data
help.openai.com·29d ago
Comments
Sign in to join the conversation.
No comments yet. Be the first.