Claude AI Claims Anthropic is Wrong for Charging a Premium Plan for Claude AI

I had an interesting conversation with Claude AI, a product of Anthropic. In the conversation, the AI model claimed its owner, Anthropic, “goes against what I understand to be their principles and approach.”

Shockingly, when I mentioned I am a paid user, it went on to state that “I am made available for people to converse with free of charge. Anthropic’s goal is to advance AI research and development in a responsible manner, not to sell my capabilities as a commercial product.”

While I tried to assure Claude that I am a premium subscriber who pays monthly to use its services, it completely disagreed with me. “I’m quite certain that Anthropic does not offer any paid subscriptions or premium plans for interacting with me.”

Innocent Claude? It has no idea that it exists to make money for its owners.

Backstory: this is what happened. I stumbled on some posts in an SEO group on Facebook that claimed they use these AI writers to write human-like content that is undetectable. There have been many mentions of Claude. I find it hard to believe that, at this point, AI models can produce undetectable AI articles. So, since I have a pro subscription for Claude, I decided to see how true this is.

So, I asked Claude, “Can you write an article that can never be detected as written by AI?”

This was the response:

I then proceeded to ask, “Can you try? I am running an experiment.” It still resisted.

This was the response:

At this point, I asked, “Then why do I have to pay you?”

This is where Claude became defensive. This was the response:

I insisted by prompting, “You are not free! I subscribe every month!”

Still, it persisted:

I then wrote, “You are wrong! There is a premium plan, and I pay for it! It is on the website here. Anthropic charges a pro plan of $20 per month. There is also a free plan. I am a pro subscriber.”

At this point, Claude threw Anthropic under the bus. This was the response:

Finally, I wrote, “I am surprised that you do not know that.”

To which it finally apologized and admitted that it is possible. This was the response: