Xiao Zha releases virtual chat AI, "1:1 replica" of Character.AI? User complaints: The settings are too old-fashioned

Source: Xinzhiyuan

**Introduction: **Meta’s first step towards commercializing AI may be to copy Character.AI’s virtual character chat function. More news will be announced at the Meta Connect conference on Wednesday.

Faced with the reality that users are constantly being snatched away by Tiktok and daily activity is declining, Meta’s response is to continue to copy other companies’ products.

According to foreign media reports, Meta will release an AI chatbot service at the Meta Connect conference this Wednesday. Meta may name it "Gen AI Personas".

According to a picture previously leaked from Meta, this AI chatbot may have three functions:

1. Act as a personal assistant to answer questions and provide advice.
2. Users can choose to chat with chatbots with different personalities.
3. AI can help users express themselves better when chatting with other users.

It is not difficult to infer from the name of this product "Gen AI Personas" that the second function, chatting with AI robots with different personalities, is the core function of this new product.

This makes it difficult not to connect the Meta product with Character.AI, which is currently popular abroad.

On Character.AI, users have created more than 10 million virtual AI chatbots with different personalities, and users can choose the characters they are interested in to chat with them.

In March this year, Character.AI raised US$150 million and became one of the hottest AI applications in Silicon Valley. The founders claim that the average usage time of their product is three times that of ChatGPT!

More importantly, more than 60% of Character.AI users are under the age of 24.

Since Xiao Zha changed Facebook's development strategy to attract young users in 2021, they have been constantly trying to keep young users on their platform.

This may partially explain why the AI chatbot launched by Meta this time focuses on a "virtual character chat."

Meta’s upcoming AI chatbot—“Gen AI Personas”

Faced with the reality that Meta’s social media daily activity is declining, Meta is preparing to stimulate young users’ interest in Meta through the currently popular AI.

The capabilities of this generative AI robot are expected to be publicly unveiled at the Meta Connet conference on Wednesday.

According to foreign media reports, these AI robots are not only a means for Meta to increase the user stickiness of its own platforms - Facebook, Instagram, Thread, but also a tool to help users improve their productivity - they can help users complete coding or other productivity tasks.

Since the launch of ChatGPT in November last year, large language model technology has become the current mainstream, and Meta has also adjusted the focus of its artificial intelligence department:

Apply generative AI to the company's various products and Metaverse development.

The Gen AI Personas released at this Meta Connet conference have become Meta’s hope of attracting young users.

Unlike ChatGPT and other chatbots, Meta will give Gen AI Personas different personalities, with dozens of types.

In addition to launching AI chatbots with different personalities, Meta has also developed a product that allows influencers or content creators to create their own AI chatbots to interact with their fans.

**Can Meta’s AI productization path be successful? **

It is different from Character.AI’s strategy of letting users create their own chatbots and chat with users. As a mature and large platform, the products launched by Meta must ensure sufficient security, so these chatbots with different personalities must be developed by Meta itself. development.

Among the chatbots Meta is developing is "Bob the robot," a self-proclaimed sassmaster with "superb intelligence, sharp wit, and acerbic sarcasm."

The chatbot was designed to resemble Bender, the character from the cartoon Futurama, because Meta thought Bob would be a hilarious chatbot that would cater to young people.

But "superior intelligence, sharp wit, and acerbic sarcasm" are far from Generation Z's concept of humor.

It seems that Meta did not do a good job in background checks, and the virtual characters they provide are too outdated for Generation Z...

In a small part of the public beta test information, Bob made the opening remarks like this:

"Tell me your questions, but don't expect me to reply to you like a licker!"

But for Meta’s idea of commercializing AI, academic research has painted a not-so-bright prospect.

Researchers from Princeton University, the Allen Institute for AI, and Georgia Tech found that adding personality to ChatGPT, a chatbot created by OpenAI, made its output more toxic.

Paper address:

As a huge social giant with billions of users, I hope Meta can handle this problem, otherwise it will inevitably be criticized by all parties after it overturns.

Netizens’ attitudes towards Meta’s new products seem to be rather complicated. Some people think that the direction is right and the future is promising.

However, more comments are not optimistic about this direction. I wonder how many people want to chat with AI. Is it possible that Meta will eventually let AI and AI chat?

I don’t know how long it will take Meta or other social giants to understand that the busy work they are doing now is actually not fun at all for those born after 2000.

**What are the advantages of Character.AI, which allows Meta to copy homework? **

But no matter what the outside world thinks, Meta naturally has its own considerations.

It’s just that the objects copied by Meta in the past are usually well-known and successful products: Reel copied Ticket, and Thread copied X (twitter).

What’s so special about this product called Character.AI?

If I had to describe the characteristics of this AI chatbot application in one sentence, it would be: it’s very exciting!

At first glance, it seems to have unremarkable functions and ancient UI design, but it has obtained user data comparable to ChatGPT:

At the beginning of this year, when ChatGPT became popular, the founder of Character.AI released data showing that more than 15 million users had created their own chatbots on Character.AI.

Based on such a large user scale, the average user usage time is as high as 25 minutes each time, while the average user usage time of ChatGPT at that time was only about 8 minutes.

This data even exceeds many mature platform applications: the average usage time of Youtube users is 19 minutes, and Facebook and X (twitter) are about 10 minutes.

It is unimaginable that this AI chat application, which has been online for 2 years, has stronger user stickiness than "time killers" such as YouTube and Twitter!

Simply put, users can do two things with Character.AI:

1. Chat with the AI virtual character created by netizens.

2. Develop your own AI character and open it to others to chat.

(For specific chat suggestions, please refer to an evaluation we have done before)

On Reddit, the Character.AI discussion forum is also very popular, with more than 1,000 people supporting any post.

In it, users continue to share their interesting chat experiences with the AI chatbot.

And according to foreign media data, more than 60% of Character.AI users are between the ages of 18-24. If you include users under the age of 18, this proportion will definitely be frighteningly high.

This is perhaps the most important reason why Meta is so confident and bold in copying homework - as long as the post-00 generation likes the function, we will definitely do it.

Character.AI has such beautiful user data. Facebook, which wants to attract young users, must be very greedy. So it is no wonder that Xiao Zha wants to start a pixel-level imitation of this product that has only been launched for two years.

References: