Talk:Character.ai

New information and updates

I’d like to suggest some updates to this article. I work for WhiteHatWiki, which was hired by the subject of this article.

@AlphaBetaGamma: Thanks very much for reviewing the previous suggestions. I'm wondering if you'd like to review these as well since you're familiar with the article.

1. Please remove the first paragraph and third paragraph of the “Software” section. Both paragraphs cite only the company website, not a reliable secondary source.

Here are those paragraphs for reference:

Built from advanced deep learning and expansive language models, the software is currently in beta and continually undergoing improvement; on 5 November 2022, conversation memory was increased by double the previous capacity so that the AI would be able to "remember" messages from farther back.

When a character sends back a response, users can rate the response from 1 to 4 stars. In addition, users can give reasons on why a certain amount of stars are chosen by clicking on one of 4 to 6 buttons. The rating predominantly affects the specific character, but also affects the behavioral selection as a whole. The user can also click the right arrow/swipe left for the AI to generate a new response, and later look through the generated messages by clicking on the left arrow.

Partly done, I've removed the first paragraph but not the second, which I've trimmed to reflect the current status of the documentation cited. Ability to rate the responses seems relevant and this is a relatively straightforward WP:ABOUTSELF. Rusalkii (talk) 06:28, 16 August 2024 (UTC)

2. Please add a new first paragraph to the Software section. This paragraph will provide information about the primary functionality of Character.ai, which is its chatbot conversation partners. This paragraph is a summary of coverage in prominent new sources including the Washington Post and Axios.

Here’s my suggestion for the new first paragraph:

Character.ai’s primary service is to provide users with the ability to have conversations with character AI chatbots that are based on fictional characters or real people (living or deceased). The responses of these characters use data the chatbots gather from the internet about a person. Some of the most popular chatbots on the site include chatbot versions of philosophers such as Friedrich Nietzsche, Socrates, and René Descartes, as well as fictional characters like Sherlock Holmes. In addition to the chat feature, users can play text-adventure games where characters guide them through scenarios.

Partly done I have removed the sentence about "most popular characters" altogether. Both sources I can see just mentions that users can chat with these people, not that these are particuarly popular. Rusalkii (talk) 06:28, 16 August 2024 (UTC)

3. Please add a new third paragraph to the Software section. This service received reliable press coverage and was the first new service the company debuted after its original beta launch.

Here’s my suggestion for the new third paragraph in the section:

The company also provides a service that allows multiple users and AI chatbot characters to converse together at once in a single chat room.

 Done, though I've added this to the end of the new first paragraph rather than as its own paragraph. I know I've made a number of substantial changes to your suggestions, feel free to ping me if you have any issues with them. Rusalkii (talk) 06:28, 16 August 2024 (UTC)

Thank you for the time you’ve spent on this review.W12SW77 (talk) 15:46, 6 June 2024 (UTC) W12SW77 (talk) 15:46, 6 June 2024 (UTC)

I'm not formally answering the request, I just wanted to note that I haven't found a really popular version of René Descartes. Friedrich Nietzsche and Socrates are indeed more popular, but they are not really "Some of the most popular chatbots on the site". If we want to focus on the most popular chatbots, we may mention the psychologist, which was featured in BBC and some other media, along with Super Mario, and perhaps Elon Musk and anime characters. So I suggest to replace the sentence "Some of the most popular chatbots on the site include chatbot versions of philosophers such as Friedrich Nietzsche, Socrates, and René Descartes." with something like:

Popular chatbots characters include anime and video game characters (e.g. Raiden Shogun, Mario), public personalities (e.g. Elon Musk), philosophers (e.g. Socrates) and therapists.

The other proposed sentences seem fine to me, and I think the request improves the article. Alenoach (talk) 13:30, 23 June 2024 (UTC)

@W12SW77 Just to let you know, I replaced the {{reflist}} template with a Reflist-talk at the bottom of this section since they're a bit easier on the eyes on talk pages. Jabster28 (talk) 11:37, 28 June 2024 (UTC)

References

Wiki Education assignment: Digital Writing

This article was the subject of a Wiki Education Foundation-supported course assignment, between 26 August 2024 and 13 December 2024. Further details are available on the course page. Student editor(s): Eithirin (article contribs). Peer reviewers: Avarice777.

— Assignment last updated by Gcutrufello (talk) 17:30, 5 November 2024 (UTC)

New fact for 'Litigation'

"After child’s trauma, chatbot maker allegedly forced mom to arbitration for $100 payout" https://arstechnica.com/tech-policy/2025/09/after-childs-trauma-chatbot-maker-allegedly-forced-mom-to-arbitration-for-100-payout/ Jaleks (talk) 07:34, 18 September 2025 (UTC)

Further controversies that should be added

I want to add these while also preserving NPOV, facts that while Character AI implemented age verification in response to controversy, it created another controversy as the company they partnered with, Persona, has had security issues that have made adult users who are over 18 falsely flagged as children decide to not upload their ID or biometric data to the site and the methods of estimating a user's age tend to skew heavily towards marking users as minors (including both poor grammar and impeccable grammar being marked as "signs the user is a minor" for example), leading many to feel the method is a form of malicious compliance with the new law to gather data for themselves rather than done for child safety assurance. One highly upvoted post on the Character AI subreddit features a middle-aged couple who are visibly obviously not teens saying that they were marked as minors (with the poster even explaining that their chats included references that a teen could not possibly know). I'm not sure how much of the above information is actually usable for the Wikipedia article (to be fair, I'm more trying to provide background info than suggesting a list of things to add to the article), but I do believe that the controversy with Persona and the method of age verification (which is similar to what a bank might require) has been a source of contention for many users and critics. ~2026-34096-86 (talk) 16:43, 5 July 2026 (UTC)

Application for ( lavozim nomi )

I am writing to exoress my strong interest in this position. ~2026-41393-40 (talk) 07:19, 26 July 2026 (UTC)