Turmeric AI reposted this
I am not sure people realize how significant this is for AI governance: #AI company Character.AI reportedly serving ~20 million users monthly, has decided to unilaterally withdraw its service to <18-year-old by November 25. Minors won't be allowed open-ended chats with AI on its platform; due to the health & #safety #risks it poses to users (and therefore the legal risks it poses to investors). This decision might come too late, since the related chatbot is alleged to have pushed a teenager to take his own life - but it is a better reaction than its competitors OpenAI & Meta, whose products are also alleged to have resulted in deaths, and whose guardrails are still being fine tuned and live-tested on society. This is a striking reminder that this AI technology and its products are the result of individuals' moral and ethical decisions. If a company can unilaterally decide to cut off a huge chunk of its user base, I am sure other companies can at the very least unilaterally decide to e.g. invest in better safeguards, prevent serious incidents, or have public health authorities approve deployment of new models to limit their own legal risks. Additional links in comment.