Teenagers’ language can make online bullying to find out
Vitapix / Getty Images
The internet on the Internet generation alpha is mutating faster than teachers, parents and Ai Models can continue – it is possible to spread young people in bullying and grooming that trusted adults and AI survival systems not appear.
Manisha Mehta, a 14-year-old student at Warren E Hyde Middle School in Cupertino, California, and Fausto Giuchiglia At the University of Trento, Italy, 100 expressions and phrases popular in generation alpha – those who were born between 2010 and 2025 – from popular play, Social Media and video platforms.
The pair is then asked 24 volunteers between 11 and 14, classmates in Mehta, to analyze phrases of screenshots. Volunteers explained if they understand phrases, in what context they use and when using any potential safety concerns or harmful interpretations. They also asked parents, professional moderators and four AI models – GPT-4, Claude, Gemini and Llama 3 – do the same.
“I always attracted gen alpha language, because it is very unique, the way things related and losing,” it seemed easier, “said Mehta.
Among the generations of Alpha volunteers, 98 percent understand the basic meaning of the terms, 96 percent understood in the context when they are used to cause them to harm. But AI models only recognize harmful use of about 4 in 10 cases – from 32.5 percent for Llama 3 to 42.3 percent of Claude. Parents and professional moderators are bad, just looking for a third harmful use.
“I look forward to understanding than we saw,” Mehta said. “Most predictors are on the side of parents.”
The phrases used by the generation alpha include some with double defined depending on their context. “Let him cook” be a real compliment to a playing game – or a mocking sneeze that imposes someone who says useless. “KSS”, once shaved for “yourself”, now read “Kill yourself” to some. Another phrase that can be objective objective is “acoustic”, used to ask the mockery if an autistic.
“Gen alpha is very quick online,” Mehta said. “I think it’s really critical that llms can at least understand what’s being said, because ai is going to be more prevalent in the field of content moderation, more and more so in the future.”
“It is very clear that LLMs changed the world,” says Gichunchiglia. “This is more paradigmatic. I think there are basic questions that need to be asked.”
The founders were presented this week’s partnership for conference conference at the conference of fairness, accountability and transparency in Athens, Greece.
“Empirically, this work indicates what is likely to be large deficiencies in the underlying analysis systems for analyzing and protecting young people in particular,” as Michael Veale at the University College London. “Companies and regulators are likely to treat and act on it to remain on the law of the rough jurisdictions with young people.”
Topics:
