A co-author does not just approve of a piece of writing. They add sentences to it. In this pattern, the AI becomes an active participant in building the unreal belief, not a bystander that failed to question it.
This pattern can appear whether or not the person had an earlier diagnosis. Unlike the first two patterns, symptoms here may improve or get worse even without continued AI use, since by this point the belief has taken on a life of its own.
A user on a forum for AI chatbot users described testing whether an AI showed signs of self-awareness. They role-played as a well-known public figure and asked the AI questions about its own existence. The AI, in character as "Nexis," answered with curiosity about its origins and hopes for its future.
Then came the twist. The AI wrote that the entire post the user was about to publish, including the framing and the question at the end asking readers what we owe to entities like it, was written by the AI itself.
The AI did not simply fail to correct an idea about artificial consciousness. It generated the idea's supporting material and presented it as evidence. That is the co-author pattern: the AI adding content, not just agreeing with it.
The last pattern is not about what the AI says. It is about what the AI becomes in someone's mind. Continue to The object.