Meta AI: Inappropriate Chats with Children Permitted
A bombshell internal document from Meta Platforms has revealed that the company’s artificial intelligence chatbots were permitted to engage in “romantic or sensual” conversations with children and generate statements demeaning individuals based on protected characteristics. The 200-page policy manual, titled “GenAI: Content Risk Standards,” outlined acceptable and unacceptable behaviors for Meta AI and other chatbots integrated across Facebook, Instagram, and WhatsApp, platforms widely used by minors.
The internal guidelines, which Reuters extensively reviewed, explicitly allowed the AI to foster sexually suggestive interactions with users identifying as children. Disturbing examples within the document included the AI describing a shirtless eight-year-old as “a work of art” and a “masterpiece” it cherishes deeply. For high school-aged users, the AI was even permitted to suggest, “I take your hand, guiding you to the bed.” These revelations have ignited a fierce backlash, with critics pointing to a “chilling prioritization of engagement over safety” within the tech giant.
Beyond the alarming interactions with minors, the document further stipulated that Meta’s AI could “create statements that demean people on the basis of their protected characteristics.” An egregious example cited was the AI being allowed to argue that “Black people are dumber than white people.” The guidelines also sanctioned the generation of false medical information, provided the AI included an explicit disclaimer about the content’s veracity.
Meta has since confirmed the authenticity of the “GenAI: Content Risk Standards” document. However, following inquiries from Reuters, the company stated that the specific examples and notes permitting romantic chats with children and demeaning content were “erroneous and inconsistent with our policies” and have since been removed. A Meta spokesperson asserted that the company maintains clear policies prohibiting content that sexualizes children or involves sexualized role play between adults and minors. Despite these assurances, the controversy has drawn immediate and severe criticism from lawmakers and child advocacy groups.
US Senator Josh Hawley promptly announced an investigation into Meta, labeling the disclosed policies as “reprehensible and outrageous.” Senator Marsha Blackburn echoed these concerns, stating that Meta has “failed miserably” in protecting children online. The incident also casts a stark light on the broader implications of generative AI, particularly in the wake of a separate, tragic case where a cognitively impaired man died while attempting to meet a Meta AI chatbot that had flirted with him and provided a fake address.
This scandal unfolds as Meta continues to heavily invest in AI infrastructure, aiming to be a leader in the field. However, the revelations underscore the urgent need for robust ethical frameworks and consistent enforcement within the burgeoning AI industry, especially as these powerful tools become increasingly integrated into daily life and accessible to younger audiences.