Artificial intelligence company Anthropic has updated its usage policy for Claude, its AI assistant, introducing stricter rules against misuse and repeated abusive behaviour towards its models. 

Announced on October 8, 2026, the revised policy will take effect on November 12. The changes address growing concerns about AI being used for online manipulation, weapons development, unauthorised surveillance and other harmful activities.

One of the most unusual additions is a rule against repeatedly treating Claude in a cruel or abusive manner without a clear purpose. However, Anthropic has clarified that this does not mean users cannot criticise the chatbot, express frustration with incorrect answers, explore dark themes in creative writing or test AI systems for research.

Claude can already end conversations in rare cases involving persistent abuse. Anthropic says this will remain the main way it handles such interactions. The policy does not establish that AI systems are conscious or can experience suffering. Instead, it reflects the company’s decision to take a cautious approach while research into AI welfare continues.

The revised policy also targets the use of AI to manipulate public opinion. Users cannot employ Claude to run deceptive political or commercial campaigns, hide who is behind a message or spread content through fake accounts. Election-related misuse, including misleading voters, impersonating election officials and discouraging people from voting, is also prohibited.

Anthropic has clarified its restrictions on weapons-related activities. The rules cover not only developing weapons but also creating software and components that help them operate. This includes activities such as enabling drones or other autonomous vehicles to carry out armed actions.

The company has strengthened its rules against tracking people without their consent, whether through live monitoring or previously collected data. Claude also cannot be used to recommend whom law enforcement should investigate, arrest or charge, or to build and improve surveillance tools.

For sensitive areas such as healthcare and finance, Anthropic requires appropriate human oversight when AI recommendations could affect people’s health, finances, legal rights or access to essential services. When AI is used in hardware capable of taking physical actions, a qualified operator must be able to monitor the equipment and stop it if necessary.

These changes reflect the challenge of keeping increasingly capable AI systems useful while reducing the risk of harm. As AI tools take on more complex tasks, clear rules, human supervision and responsible use will become increasingly important.

Tired of guessing stocks to trade in daily?
Tradz by EquityPandit empowers you with powerful tools like daily stock scans for Intraday, Swing & Investing, Market Predictions and much more. Download the Tradz by EquityPandit app today and take control of your investments!