The renowned AI startup Character AI recently made headlines with news of its financial risks, potentially leading to an acquisition by Google or Meta. What has caused a unicorn valued at $1 billion to face bankruptcy risks within just two years? In this article, we will analyze the reasons behind this from perspectives such as the company’s product design, corporate philosophy, and founder backgrounds.
What is Character AI?
Character AI is an AI company founded in 2022 by Noam Shazeer, formerly of Google Brain. The company developed an AI for roleplaying (LLM), known as Character.ai. Its main purpose is to allow users to create various AI characters and interact with them.
Its primary use case is AI roleplaying, similar to NsfwGPT.ai. Its official website features numerous well-known characters from business stars to anime characters and movie stars, which is a major attraction for the company.
Why is Character AI facing financing difficulties?
Previously, Character AI received $160 million in investment from A16Z, catapulting it to a valuation of $1 billion. However, it is now facing financial difficulties that threaten its funding chain, necessitating an acquisition. Behind all this, I see contradictions between Character AI’s team, product design, and corporate philosophy. The current situation for c.ai (Character AI’s abbreviation) is something that could have been anticipated.
“We have the strictest NSFW filter”
A spokesperson for Character AI said in a statement, “Among all consumer AI platforms, we have the strictest anti-NSFW policy.” This stance is fatal. I believe that among all types of AI products, such as those for roleplay or AI companionship, Character AI should be the least opposed to NSFW content. However, they have the strictest NSFW filter among all AI products. This could be due to two factors: firstly, their team background, coming from large tech companies like Google, where founders may hold more noble views, deeming adult content unacceptable and thus strictly opposing NSFW. Secondly, pressure from investors who may fear the various risks associated with NSFW content, hence not allowing the c.ai team to open NSFW products. However, major companies now recognize that NSFW content is difficult to avoid and not necessarily immoral in adult entertainment, gradually opening up NSFW checks, such as Meta’s new open-source LLM llama3’s NSFW filter much looser than the previous llama2. OpenAI also clearly stated that it is considering opening up adult content creation, but character.ai runs counter to the mainstream.
Users feel c.ai models are becoming more “impersonal”
Recently, we can see many complaints on Character AI’s official subreddit, where people feel that c.ai’s AI characters are becoming increasingly impersonal, with responses becoming emotionless. Although the c.ai team denies modifying their models, it is evident from most user complaints that there must be some issue with their models. I think it might be because they further filtered NSFW content, leading to extensive neutering of the new training dataset. It is well known that LLMs require large amounts of textual data for training, especially roleplay applications like Character AI requiring extensive literary work datasets, in which sexual or suggestive content frequently appears, c.ai may have deleted part of the data, which causes the fine-tuned model to only learn pale language.
Grand vision and user demands are at odds.
At its inception, C.AI had a very grand vision; they aimed for general artificial intelligence. It’s clear that this team harbors immense ambitions. Consequently, they disdain explicit content, even strongly reject it. However, sexual needs are inherent to every individual. The foundation of products like C.AI, which involve roleplay or AI companionship, lies in human emotional needs. While other AI tools may have NSFW filters, I believe Character.AI should not. Yet, they are among the most vehemently opposed to NSFW content in the entire market. One might say their lofty vision and illustrious founder backgrounds are diametrically opposed to the foundation of their products.
AI companions that oppose pornography have no future.
Imagine what AI companions will be like in 10 years. They will have more human-like emotions, more human-like tones, and perhaps even human-like bodies with perfect shapes. They could be with you every day, and some might even engage in intimate relations. However, C.AI is so averse to NSFW content that its robots refuse to discuss any adult topics. How many people would want such a companion? It would only lead to greater loneliness. A perfect AI must encompass all human emotions and desires; otherwise, it becomes unreal.
Using NsfwGPT.ai as an alternative to Character AI.
NsfwGPT.ai is also built on advanced AI technology, and importantly, it does not impose filters on explicit content. We are an open and unrestricted platform that allows any legally and ethically acceptable topics. We do not view explicit content as unethical. Come experience the limitless AI roleplay chatbot with NsfwGPT.ai now.
