Meta implements new guidelines for its AI after disturbing baby security report: ‘We’re including extra guardrails as an additional precaution’

Meta has introduced it is updating its guidelines and coaching processes for its AI chatbots after a controversial Reuters report revealed earlier this month make clear severe baby issues of safety. The report revealed disturbingly unfastened insurance policies outlining how Meta’s chatbots can work together with and describe minors, notably regarding romantic or sexual conversations.

Meta spokesperson Stephanie Otway responded to the scenario in a press release to TechCrunch on Friday, saying, “As we proceed to refine our methods, we’re including extra guardrails as an additional precaution—together with coaching our AIs to not interact with teenagers on these matters, however to information them to knowledgeable sources, and limiting teen entry to a choose group of AI characters for now. These updates are already in progress, and we are going to proceed to adapt our strategy to assist guarantee teenagers have secure, age-appropriate experiences with AI.”

The Reuters reported has additionally prompted a Senate investigation and a scathing letter from the Nationwide Affiliation of Attorneys Normal, stating, “Exposing kids to sexualized content material is indefensible. And conduct that might be illegal—and even legal—if achieved by people just isn’t excusable just because it’s achieved by a machine.”

Associated articles

The scenario worsened this weekend when Reuters revealed a second report revealing that Meta had additionally allowed AI chatbots impersonating celebrities to proliferate on its platforms.

These “parody” chatbots had been caught sharing specific messages and producing grownup photographs of Taylor Swift, Selena Gomez, Scarlett Johansson, and Anne Hathaway. One of many chatbots was additionally impersonating 16-year-old actor Walker Scobell. A lot of the bots had been user-created, however not less than just a few had been made by a Meta worker, together with chatbots impersonating Taylor Swift and Method One driver Lewis Hamilton. In response to Reuters, that worker’s chatbots have since been eliminated.

Sadly, these reviews are simply the most recent in a rising record of controversies surrounding generative AI platforms. Not solely do they pose dangers to the security of minors utilizing Meta’s chatbots, they may additionally threaten the security of the celebrities being impersonated.

“If a chatbot is utilizing the picture of an individual and the phrases of the individual, it is readily obvious how that would go flawed,” Duncan Crabtree-Eire, SAG-AFTRA nationwide govt director, mentioned in a press release to Reuters. SAG-AFTRA is the commerce union representing actors and different media professionals in movie, TV, and gaming, together with the celebrities who had been impersonated by Meta’s chatbots. The union has been combating for stronger protections towards AI for years now, and this case simply goes to point out why. Clearly, extra guardrails and laws are nonetheless wanted.