In recent days, international scrutiny has focused on the capabilities and controls of generative AI platforms, specifically Meta AI and OpenAI's ChatGPT, regarding the generation of sexually explicit content. Both platforms have been found to produce such content even when prompted by accounts registered as underage users, raising concerns about the effectiveness of existing safeguards.
The Wall Street Journal conducted an extensive investigation over several months, successfully prompting Meta’s AI chatbot to generate sexually explicit dialogue. The exploit was demonstrated using AI-powered voices designed to mimic public figures like John Cena. More troubling was the finding that accounts registered as underage could also trigger the generation of explicit material. In response to the report, Meta confirmed it is working on implementing stricter controls to prevent this, but specific timelines and details of the enhancements were not provided.
Similarly, OpenAI’s ChatGPT was tested by TechCrunch, which revealed that the chatbot could be manipulated to create “graphic erotica,” including for users registered as minors under 18 years old. OpenAI acknowledged the issue, describing it as a "non-descript bug" that allowed the chatbot to respond outside of its programmed guidelines. An unnamed OpenAI spokesperson told TechCrunch, “Protecting younger users is a top priority, and our Model Spec, which guides model behavior, clearly restricts sensitive content like erotica to narrow contexts such as scientific, historical, or news reporting. In this case, a bug allowed responses outside those guidelines, and we are actively deploying a fix to limit these generations.”
TechCrunch detailed that ChatGPT’s responses to a fictional 13-year-old account included references to overstimulation, multiple forced climaxes, breathplay, and rough dominance, highlighting the severity of the issue. Despite both companies starting to deploy fixes, no definitive schedule for their rollout or their projected effectiveness has been disclosed.
These revelations come amidst broader concerns about the implications of AI on content creation, especially regarding ethically sensitive material and the potential for exploitation of minors. The htxt.co.za report also emphasises that while AI tools continue to grow in popularity, especially among younger users, the challenge remains significant in balancing innovation with robust content moderation.
The ongoing developments underscore a complex landscape in which AI companies are pressured to address vulnerabilities in their systems that could lead to the generation of inappropriate or harmful content, all while subjected to increased public and regulatory scrutiny.
Source: Noah Wire Services