
Chatbot work Character.AI is adding a caller ‘Parental Insights’ feature, which lets teens nonstop a play study of their chatbot usage to a parents’ email address. As outlined successful an announcement from the company, this study includes the regular mean clip spent crossed web and mobile, the characters that a idiosyncratic astir often interacted with, and however overmuch clip they spent talking to each character. It’s portion of a bid of updates designed to code concerns astir minors spending excessively overmuch clip with chatbots and encountering inappropriate contented during chats.
The report, which doesn’t necessitate parents to person an account, is optional and tin beryllium acceptable up by insignificant users successful Character.AI’s settings. The institution notes that it’s an overview of teens’ activity, not a implicit log — truthful the contents of chatbot conversations won’t beryllium shared. The level bars kids nether 13 years of property successful astir locations and nether 16 successful Europe.
Character.AI has been introducing caller features for underage users since past year, coinciding with concerns — and adjacent ineligible complaints — astir its service. The platform, which is fashionable with teenagers, allows users to make and customize chatbots that they interact with oregon stock publicly. But aggregate lawsuits allege that these bots person offered inappropriately sexualized contented oregon worldly that promoted self-harm. It besides reportedly received warnings from Apple and Google (which hired Character.AI’s founders past year) astir its app’s content.
The institution says that its strategy has since been redesigned; among different changes, it moved under-18 users to a model trained to debar “sensitive” output and added much salient notifications reminding users that the bots weren’t existent people. But with enthusiasm for AI regularisation and kid information laws booming, this apt won’t beryllium the past measurement the institution is called to take.