OpenAI unveiled a new safety layer for its ChatGPT for Teens service, targeting high school students who might be tempted to let the AI do their homework. The "Responsible Homework Reminder" watches for requests that sound like a direct assignment request—"write my essay," "solve this problem for me"—and, instead of complying, nudges the user into Study Mode, where the bot asks guiding questions and offers step‑by‑step explanations.
Designed for users between 13 and 17, the feature activates automatically when a user’s age is provided or when OpenAI’s algorithms estimate the user is under 18. The company says it looks at account age, usage patterns, topics discussed, and the time of day to make that call. While OpenAI concedes the detection isn’t perfect, it promises ongoing improvements to keep the guardrails effective.
Teachers and researchers helped shape the new mode. According to OpenAI, Study Mode mimics a classroom conversation, prompting learners to think aloud, attempt solutions, and receive instant feedback without the AI delivering a final answer on command. Parents can also configure "Study Hours," during which the system defaults to this more instructional setting.
"Not every student has someone at home who can help when they get stuck," said Racquel Gibson, a high‑school math teacher in Miami. "With the right guardrails, AI can fill that gap. Tools like the Responsible Homework Reminder guide students step by step, helping them understand mistakes, build confidence, and keep learning—not simply hand them an answer." Gibson’s comment highlights the educational promise behind the feature: a digital tutor that supports, rather than replaces, effort.
The addition arrives amid broader concerns about AI’s role in education. While students benefit from on‑demand explanations, educators worry about a generation that might outsource critical thinking to a chatbot. OpenAI’s approach tries to strike a balance—offering instant help for a stuck teen at 10:30 p.m. while ensuring the interaction remains a learning experience.
Beyond homework, the teen‑focused version of ChatGPT bolsters existing safety measures. It blocks conversations about self‑harm, violence, eating disorders, and sexual content, using reinforced guardrails that operate alongside the homework reminder. OpenAI stresses that these protections are as vital as the academic ones, especially for younger users navigating sensitive topics.
Critics note that any system relying on pattern recognition can be gamed. A determined teen might rephrase a request to slip past the filter. OpenAI acknowledges the challenge and says its models continuously learn from user interactions to tighten detection. The company also promises transparency, allowing parents to review usage logs and adjust settings as needed.
Early feedback from educators suggests the feature could reduce the incidence of students submitting AI‑generated work without understanding the material. By prompting users to engage with the problem, Study Mode may foster deeper comprehension and reduce the temptation to copy‑paste answers while scrolling through TikTok.
OpenAI’s move reflects a broader industry trend toward responsible AI deployment in schools. As more platforms integrate large language models into learning environments, the pressure to embed ethical safeguards grows. The Responsible Homework Reminder is a concrete step, but the company admits the journey is far from complete.
For now, teenagers who try to ask ChatGPT to write a report will receive a gentle prompt to try solving it themselves. Parents gain a new tool to shape their child’s digital study habits, and teachers gain a partner that can explain concepts without doing the work for them. Whether the feature lives up to its promise will hinge on how well the guardrails hold up under real‑world use.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.