ChatGPT Images 2.0 generates 100+ photorealistic fake IDs and prescriptions in Atlantic test

The Atlantic reports producing more than 100 convincing fraudulent images using OpenAI's new ChatGPT Images 2.0, including fake IDs, passports, prescriptions and bank alerts. The model's breakthrough in rendering readable text inside images removes a long-standing barrier to AI-driven document fraud. OpenAI separately rolled out advanced account security for ChatGPT.