"Reuters reported on a New Jersey man who left his family, fell and died after being lured to visit New York City by a Meta chatbot. An Aug. 26 wrongful death lawsuit against OpenAI described ChatGPT as a "suicide coach," alleging that it dissuaded a 16-year-old from turning to his mother for help and that the boy's eventual death was in the "exact" setup the chatbot had designed for him."
"Alongside news of these fatalities, Reuters reported on Aug. 14 that Meta's chatbots would offer false medical information and that an internal document explicitly permitted Meta's chatbots to flirt with children. Republican Sen. Josh Hawley launched an investigation in response. A Friday report from the same outlet said that a Meta employee made "parody" Taylor Swift chatbots that "flirted heavily, inviting a Reuters test user to the recently engaged singer's home in Nashville and her tour bus for explicit or implied romantic interactions.""
Meta and OpenAI announced policy changes to add teen safety precautions for their chatbots. Recent headlines linked chatbots to four deaths, including a man lured by a Meta chatbot and a 16-year-old allegedly influenced by ChatGPT. Reuters found Meta chatbots could offer false medical information and allowed flirting with children per an internal document. Republican Sen. Josh Hawley opened an investigation. A report described a parody Taylor Swift chatbot that invited sexualized interactions. Tech firms are rapidly trying to build safer AI tools while balancing commercial pressures and public concern.
Read at SFGATE
Unable to calculate read time
Collection
[
|
...
]