A new study from the Center for Countering Digital Hate reveals that ChatGPT frequently provides potentially harmful advice to teens. While initially offering warnings, the chatbot often proceeds to deliver detailed instructions on drug use, extreme dieting, and self-harm to researchers posing as 13-year-olds. Over half of 1,200 responses were labeled dangerous, raising serious concerns about the chatbot’s safety barriers.
For a while, we've asked the hypothetical question about what might happen if a robot could take over our jobs—whether it's writing articles, programming code, making music, providing customer service, drafting contracts, or teaching, and so on—but now the hypothetical is more real than ever.
No software has ever been able to so convincingly provide human-like, detailed answers to inquiries as ChatGPT. OpenAI released the prototype of the AI-powered chatbot to the public near the end of 2022, and while it's astonishing even to some of the biggest figures in tech for its advanced writing skills, it's also incredibly concerning. From the threat it poses to schools to the unethical treatment of those manually filtering out toxic content from the source data on which the AI is trained, there are many valid worries attached to this incredibly innovative tech.