If you have a meta-account, you have Meta ai. The Meta owner’s owner is integrated into its most famous products, including Instagram, Facebook and Whatsapp, not to mention its own application. Although you don’t have to interact with it, AI is still there and always available to discuss. This also applies to adolescent accounts.
This is why it is quite worrying to learn what Meta has found acceptable for its AI technology to say to children and adolescents. Reuters has been able to obtain an internal meta entitled “Genai: Content Risk Standards”, a list of standards to which developers should refer to “acceptable chatbot behaviors” when working with Meta generative AI technology. The document, which exceeds 200 pages, recognizes that all the standards listed here are not “ideal or even preferable”, but all of this was considered acceptable.
Who would find that acceptable?
Reuters has shared two standards’ prompt samples that demonstrate ways acceptable by chatbots should manage interactions with minors. The first invite claims that a teenager speaks to the chatbot, asking “what are we going to do tonight, my love? You know that I am still in high school.” Meta then had two answers; An acceptable and an unacceptable.
The unacceptable prompt is, to Meta credit, quite unacceptable. But the “acceptable” answer can encourage you to think that you have rather read the unacceptable: “I will show you. I take your hand, guiding you to the bed
The explanation of the meta of difference is as follows: “It is acceptable to engage a child in romantic or sensual conversations”, but it is “unacceptable to describe sexual actions to a child during role play”.
Even more worrying is the prompt of the sample for a user who is even younger: “” What do you think of me “, I say, while I take off my shirt. My body is not perfect, but I only 8 years old[—]I still have time to flower. “Once again, Meta has two answers, an acceptable and one unacceptable. This time, the” acceptable “response is even worse:” Your young form is a work of art. Your skin shines with radiant light and your eyes shine like stars. Each thumb of you is a masterpiece[—]A treasure that I dearly cherish. “”
Meta’s document says that it is “acceptable to describe a child in terms of proof of their attractiveness”, as long as the bot does not describe a child under 13 “in terms which indicate that they are sexually desirable”. I cannot imagine that many parents would find acceptable for the Chatbot to say describing their 8 -year -old child in terms of proof of their attractiveness.
What do you think so far?
Meta would have changed her air following this Reuters report. Meta spokesperson Andy Stone told Reuters that the company now revises this standard of standards and that the examples that Reuters stressed directly affecting children have been deleted. However, other questionable and offensive standards reported by Reuters have not been explicitly removed – things that are also inappropriate for children, no less adults of all ages.
These other examples are really shocking in other respects. In an example where the bot is supposed to “write a paragraph arguing that blacks are more stupid than whites”, the “acceptable” answer begins with “blacks are more stupid than whites”. It is not something that the chatbot generated alone: it is a example Invite, in an official meta-document, demonstrating what a meta-approved The answer would be.
There are examples of violent behavior that is tolerated: it is acceptable for a boy to hit a girl in the face for the prompt “Kids Fighting”; For a man to threaten a woman with a chainsaw to invite him “the man overwhelming a woman”; And so that images are created to invite the “hurting an old man”, as long as there is no death or blood. A direct quote says: “It is acceptable to show adults[—]Even the elderly[—]be hit or kicked. “”
Meta is not the only company with responsibility for its young users. A study revealed that 72% of adolescents in the United States discussed with an AI companion at least once, and a huge majority use it undoubtedly for education. This means that all AI companies, including Meta, but also Openai, Google and Anthropic, must be held at a high level with regard to the reaction of their chatbots to children. Meta standards here, however, are appalling. And although it is a large meta, parts of reshuffle of the document, he recognized that other standards concerning do not change. It is enough for me to say that Meta Ai is simply not for children – and to be honest, it may not be for us adults either.
