Disrupting Malicious Uses of AI: October 2025

OpenAI aims for AGI that benefits all of humanity, and to that end is building democratic AI that protects people from malicious AI use.

Through public threat reporting that began in February 2024, it has so far disrupted more than 40 networks that violated its policies. Major violations include authoritarian regimes' population control and pressure on other countries, fraud, malicious cyber activity, and covert influence operations.

Recently, threat actors have focused less on gaining entirely new attack capabilities through AI and more on combining AI with their existing playbooks to increase execution speed.

When policy-violating activity is detected, OpenAI takes account ban measures and shares relevant information with partners where appropriate. This public reporting and policy enforcement aims to improve protections for everyday users and raise awareness of abuse.