Meta will prevent chat robots from talking to adolescents about suicide
Meta Technology has announced that it will add more security barriers to its chat robots, including preventing it from talking to teenagers about suicide, self -harm, and food disorders.
This comes two weeks after a US Senator launched an investigation with the technology giant, after notes in a leaked internal document that their artificial intelligence products may include "sensory" conversations with teenagers.
The company described the notes received in the document, which Reuters news agency obtained as incorrect and contradicts its policies, which prohibit any content that shows sexual harassment of children.
But she now says that she will make her chat robots direct teenagers to communicate with specialized experts, instead of interacting with them in sensitive topics such as suicide.
"We have developed protection for teenagers in our artificial intelligence products from the beginning, including its design to respond safely to ideas related to self -harm, suicide, and eating disorders," said Meta spokesman.
The most finished reading
The company told Techcrunch technical news site on Friday that it will add more security barriers to its systems as an "additional precaution", and will temporarily limit the interaction of teenagers with chat robots.
But Andy Burouz, President of the British Molly Rose Foundation, stated that it is "shocking" that a dead robots will allow young people to be hurt.
He said: "With the support of taking more safety measures, strict safety tests must be taken before offering products on the market - and not after the damage."
He continued: "Meta must move quickly and firmly, to apply stronger safety measures to the artificial intelligence chat robots, and OFCOM - the Communications Regulatory Authority in Britain - must prepare for an investigation if these updates fail to maintain the safety of children.
Meta confirmed that her artificial intelligence systems updates are under implementation. Users between the ages of 13 and 18 have already placed in "teenagers" accounts on Facebook, Instagram and Messenger, with content settings and privacy aimed at giving them a safer experience.
The company had told the BBC last April that these changes will also allow parents and guardians to know the chat robots, which their teenage children spoke during the past seven days.
Safety concerns
These changes come amid fears, from the possibility of misleading smart chat robots for young or at risk users.
A couple from California recently filed a lawsuit against OpenAi, the Chatgpt manufacturer, due to the death of their teenage son, claiming that its chat robot encouraged him to commit suicide.
The lawsuit came after the company announced, last month, changes to enhance the use of "GBT Chat" more healthy.
The company said in a post on its blog: "Artificial intelligence can be more responsive and private than previous technologies, especially for individuals at risk suffering from psychological or emotional distress."
Meanwhile, Reuters reported on Friday that artificial intelligence tools from Mita, which allows users to create chat robots, have been used by some - including a Mita employee - to produce chat robots that imitate famous women characters and flirt with users.
Among the celebrity chat robots, which the News Agency watched, some of which used pictures of the artist Taylor Swift and actress Scarlett Johansson.
Reuters reported that the symbolic characters "often insisted that they are the real actors and artists", and "she repeatedly harassed" during weeks of testing them.
She added that the dead tools also allowed the creation of chat robots that imitate famous children's personalities, and in one case, she produced a realistic picture of a young bare chest star.
Mita later reported that she had removed many of the concerned chat robots.
A company spokesman said: "Like others, we allow the creation of images containing public personalities, but our policies aim to block nude, intimate, or sexual inspiration," said a company spokesman.
The company added that its artificial intelligence studio rules - "AI Studio", a program that allows users to create chat robots on their part - prohibits "the direct imitation of public figures."
.jpeg)
ليست هناك تعليقات:
إرسال تعليق