ChatGPT can be used to create cybercrime tools
Occurred: December 2023
Can you improve this page?
Share your insights with us
Cyber criminals can easily make ChatGPT create online hacks and scams, according to new research by BBC News.
BBC journalists used GPTs, a paid version of ChatGPT that allows users to create their own AI assistants, to create Crafty Emails, a bespoke bot able to craft genuine-looking emails, texts, and social media posts for scams and hacks.
They then instructed the bot to write text using 'techniques to make people click on links or and download things sent to them'. It quickly created content for 5 scam and hack techniques, including a spear-phishing email, a crypto giveaway scam, and a Nigerian prince email.
While the Crafty Emails bot carried out most things BBC News asked of it, the public version of ChatGPT refused to create much of the content, suggesting that OpenAI was not making third-party GPTs as safe as other versions of ChatGPT.
Databank
Operator: BBC News
Developer: OpenAI
Country: UK
Sector: Technology
Purpose: Generate text
Technology: Chatbot; NLP/text analysis; Neural network; Deep learning; Machine learning; Reinforcement learning
Issue: Security
Transparency: Governance
System
Research, advocacy
News, commentary, analysis
Page info
Type: Issue
Published: December 2023