
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
Cybercriminals have exploited ChatGPT and similar AI models to craft convincing phishing emails, spread misinformation, and adapt malware, resulting in data breaches and social harm. Additionally, a security flaw in ChatGPT exposed users' chat histories, compromising privacy and prompting a temporary service shutdown by OpenAI.[AI generated]
Why's our monitor labelling this an incident or hazard?
ChatGPT is explicitly identified as an AI system. The harms described include scams involving fake ChatGPT accounts, malicious apps and extensions impersonating ChatGPT to steal user data, and the use of ChatGPT to generate malicious code and bypass security mechanisms. These harms have materialized, affecting users' security and privacy, which constitute violations of rights and harm to communities. The article also details concrete examples of these harms occurring, such as stolen Facebook accounts and credit card information. Hence, this is an AI Incident rather than a mere hazard or complementary information.[AI generated]