Anthropic Updates Claude Policy and Invests $150 Million
Starting November 12, 2026, Anthropic will have an updated Usage Policy for Claude. It now includes a consolidated section against deceptive campaigns and artificial activity. This includes, among other things, concealing senders and artificially amplifying content through fake accounts or posts.
At first glance, this sounds like a set of rules for the grand political arena, but according to the update, it also affects smaller-scale digital projects. Anyone running their own website, forum, or AI-powered service should take a closer look at the section. Especially when Claude is used in areas where errors can be more than merely annoying and can have quite serious consequences for those affected.
Humans remain in control of high-risk applications
For high-risk applications, Anthropic continues to require a qualified human oversight authority. This person or entity must not merely be allowed to watch but must also have the authority to make changes. A human signature on an automated process is therefore not enough if it no longer has any influence on the outcome.
The rule applies to applications that may affect health, legal rights, finances, livelihoods, or access to essential services. That is a broad list, but not a particularly hidden one. Claude therefore must not simply act as an invisible decision-maker in these areas.
In addition, affected individuals must be informed that AI was used. For operators of such services, this means that the use of AI should not disappear behind a human-sounding interface. Especially when a system touches on health, money, legal matters, or a person's livelihood, it should remain clear that Claude is involved.
The new consolidation thus brings together two different problem areas. On the one hand, it concerns deceptive campaigns, concealed senders, and artificially inflated content. On the other hand, it concerns applications where people must receive qualified oversight and a clear notice that AI is being used. The common denominator is easy to identify: Claude should be used neither as an invisible amplifier nor as a secret decision-maker.
Claude is set to enter US government agencies and research with $150 million
Alongside the policy update, Anthropic is announcing $150 million over three years for the US Genesis Mission initiative. The money is intended to make Claude available to more than 15 participating federal agencies. Those named include NASA, the National Institutes of Health, and the National Science Foundation.
The package consists of more than just access to Claude. It also includes Claude Code and API credits for several hundred research projects. Training, onboarding, and technical support are also included. In other words, Anthropic does not merely want to activate a few accounts and then send users on their way with a friendly “Good luck.”
The announcement thus establishes a connection between public research and several of the company's offerings. In addition to Claude itself, Claude Code will also be used, while API credits are intended to enable integration into research projects. The available material does not describe in greater detail exactly how the individual projects will use the offerings.
The timing is notable: While Anthropic wants to expand the use of Claude in government agencies and research projects, the company is also drawing clearer boundaries around its use. No artificially amplified content through fake accounts, no concealed senders, and continued human oversight in high-risk applications, along with notice to those affected. Claude is therefore supposed to become bigger, but not more invisible.
Sources
See more of us in Google search results: add us as a preferred source.
Add as preferred source on GoogleWritten by
Daniel Böhmel has been developing software for over 20 years and has been self-employed since 2007. He started out with interactive web design back in the Flash era; later, he built his own online projects, such as a fitness forum and a supplement store, including the editorial content. Today, his…
View the profile of Daniel BöhmelParticipate now!
Don’t have an account yet? Register yourself now and be a part of our community!