The latest usage policy update introduces new regulations concerning propaganda campaigns, surveillance, and weapon development.
Anthropic is modifying its usage policy for the first time in over a year to address high-risk misuse cases, including election interference, weaponry, surveillance, and financial issues. A significant focus is the prohibition of "sustained and needless abusive or cruel behavior" directed at Claude.
Last August, the company announced it could end conversations with persistently harmful users. The new update maintains termination as the primary enforcement method. Anthropic notes this rule applies only to extreme cases of cruel behavior with no discernible purpose, excluding common frustration or dark creative themes.
Anthropic has consolidated scattered restrictions into a ban on deceptive commercial or political campaigns, targeting AI-generated propaganda. This includes obscuring the origin of messages and amplifying content via fake accounts. Additionally, the election section forbids voter deception, misinformation, and impersonation of officials.
While weapons development was previously banned, Anthropic has observed attempts to use Claude for guidance and control software. The updated policy now explicitly bans actions like arming drones and building weapon components. Furthermore, the company has clarified that Claude cannot be used to recommend individuals for investigation or arrest.
The company states these rules may be modified for specific government contracts. A new safety rule requires a qualified operator to observe and stop autonomous hardware capable of physical injury. Despite exploring the potential consciousness of models, Anthropic maintains that they do not know if the models are conscious.
A free daily digest of the news that matters most.
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.