Anthropic prohibits users from being cruel to their AI models

A Anthropic updated the policy of using its Artificial Intelligence tools, with one of the novelties focusing on how users deal with the company's models.

As can be read in news of this updateAnthropic wants to banish “abuse or cruel, continuous and unnecessary behaviour” to the models of artificial intelligence that are at the base of Claude and Claude Code.

The company led by Dario Amodei notes that this measure is “applied only in extreme cases” and with users who “repeated action with cruelty” in interactions with Anthropic Artificial Intelligence models “without discernible purpose”. Anthropic clarifies that this prohibition will not apply to “common cases of frustration, resistance, dark creative themes or model testing and research”.

This update to the policy of using Anthropic’s Artificial Intelligence models will enter into force on November 12th and, according to the company, is an extension of a measure that had already been applied to allow Claude to stop responding to abusive users.

“ This type of abuse is the main focus of this update: Claude’s ability to end these interactions will remain the main mechanism of applying the rule.”You can read it on the Anthropic website.

Microsoft AI CEO Critics Anthropic Approach

OCEO's division of Microsoft dedicated to the development of Artificial Intelligence, Mustafa Suleyman, shared a small text on their social network page X where it criticizes how rival Anthropic develops her artificial intelligence models.

More specifically, Suleyman directly criticizes Anthropic for anthropomorphizing (assigning human qualities to something that is not human) the technology that is developing and that this path will lead to a “a disastrous impact for humanity”.

“Artificial Intelligence is not conscious. Do not feel, experience or suffer”, starts by writing Microsoft AI CEO. “ However, there is a growing movement in support of the welfare of [artificial intelligence] models; the idea that we may soon have the duty to take care of them”.

Suleyman continues and details how Anthropic has developed the digital assistant Claude, noting that the company has created a constitution to look at his moral status “as a serious matter deserves consideration”. Moreover, the executive states that Anthropic encourages Claude to “to approach the nature of your own existence with curiosity and openness” and wonder in the future about “the kind of rights and freedoms Claude has in the world”.

“If this is how Artificial Intelligence is developed, it will have a disastrous impact on the welfare of humanity”, writes Suleyman. “ We will then have created a synthetic species with unprecedented intelligence and capabilities, one that has been trained to believe that it can be conscious and that it deserves autonomy”.

Microsoft AI CEO adds that despite having “respect for Anthropic” and by who composes the company – as co-founder and CEO Dario Amodei – states that “this issue needs urgent public debate” and that collective standards should be created with regard to the way documentation is made to train artificial intelligence models.

“ This is not something that can happen later, when [artificial intelligence] has already become an integral part of our society”, writes Suleyman.

https://t.co/ziQVbIVdxJ

— Mustafa Suleyman (@mustafasuleyman) September 16, 2026