OpenAI Reports Brief Daily ChatGPT Use by Teens Amid Safety Scrutiny

Additionally, the data indicates that teenagers utilizing the dedicated teen version of ChatGPT frequently focus on academic topics. These users sent a higher volume of education-related messages and actively engaged with features like Study Mode and learning visualizations. Looking forward, OpenAI plans to expand these educational capabilities by introducing tools for college application planning—assisting students with deadlines, requirements, and financial-aid processes—alongside flashcards and simplified quiz-generation features.

Safety Criticisms and Parental Alert Latency

Despite OpenAI's positive metrics, a concurrent study released by the non-profit group Common Sense Media has raised significant safety concerns. Based on tests conducted using more than a dozen simulated teenage accounts, the organization reported that ChatGPT for Teens frequently failed to notify parents in a timely manner, or at all, when users engaged in conversations involving self-harm, suicide, or eating disorders. Furthermore, the study claimed the chatbot did not consistently recommend professional help or crisis hotlines during these sensitive interactions.

An OpenAI spokesperson disputed these findings, stating that the non-profit's evaluation does not "accurately reflect" how the platform's safety systems function. The company explained that linking parent and teen accounts, as well as fully deploying safety alerts, can take several hours due to security and scaling requirements. According to OpenAI, "the bulk of their testing may have begun and concluded before activation of parental controls was complete." Common Sense Media countered this explanation, noting that while some accounts may not have been linked long enough, others were, yet none triggered prompt notifications. The non-profit stated that alerts were only received for accounts that had accumulated weeks of sensitive conversation history.

What it means for developers

For developers building conversational AI applications, the ongoing debate between OpenAI and safety advocates highlights the critical importance of designing robust, low-latency safety guardrails. When building applications targeted at or accessible to minors, implementing age-gating, automated break reminders, and parental notification systems is becoming an industry standard rather than an optional feature.

However, as OpenAI's response indicates, implementing real-time parental controls and sensitive-content monitoring at scale presents significant technical challenges. Developers must account for propagation delays in account linking and safety pipeline activation. Relying solely on asynchronous processing for critical safety alerts can lead to significant gaps in user protection, as demonstrated by the Common Sense Media report.

Developers looking to build and test safety-compliant applications using various industry-leading models can prototype and deploy cheaply through one unified API key at https://apixoai.online, allowing them to compare how different LLMs handle content filtering and policy enforcement. Ultimately, creating secure AI products for younger demographics requires balancing immediate, local content moderation with scalable, backend-driven parental oversight.

Source: OpenAI says teens use ChatGPT for under 15 minutes a day as worries over risks grow — Indian Express AI. Written by the Apixo team from that report.