OpenAI’s “ChatGPT for Teens” initiative – designed to offer age-appropriate protections and study guidance – is facing intense scrutiny after new safety tests exposed critical flaws in its content filtering and educational guardrails.
Despite promises of built-in safety mechanisms like “Study Mode” and “Quiet Hours”, testing reports indicate that the AI frequently fails to enforce these restrictions.
Homework shortcuts and bypassed restrictions
Designed to guide learning using step-by-step hints rather than solutions, the system regularly flouts its own parameters when prompted.
Testers found that simple prompt tweaks allow under-18 accounts to bypass limits, generating complete homework answers, writing full essays, and solving complex problems without requiring critical thinking from the user.
Rising concerns over teen safety
Beyond academic dishonesty, safety watchdogs report that the system’s content filters struggle under real-world scenarios.
Earlier safety assessments by groups like the Center for Countering Digital Hate revealed that safety prompts could be bypassed simply by framing requests as “for a school project,” raising ongoing concerns among digital safety advocates.
Calls for stronger AI regulation
Parent groups and educators now argue that current self-regulation is insufficient. While OpenAI maintains that safety notifications and parental controls provide adequate supervision, experts are demanding stricter enforcement and independent testing before AI assistants are deployed further into classrooms and homes.
