Researchers, mainly from Google’s DeepMind, have found that OpenAI’s large language models, including closed-source model ChatGPT, contain significant amounts of private identifiable information (PII). Using an innovative prompt, the team could extract verbatim data memorised from training, such as email addresses and phone numbers. Approximately 16.9% of generations tested included memorised PII. The method could, given adequate funds, potentially unearth gigabytes of training data.
Group-IB’s Threat Intelligence and Defence Centre Equip Undergraduates with Sophisticated Cybersecurity Technologies to Boost Threat Analysis and Enhance Cyber Resilience for Campus Start-ups
Hey there from the heart of the San Francisco Bay Area! It’s an absolute pleasure to have you back again for our chat on some
