WASHINGTON, D.C. — OpenAI’s new “ChatGPT for Teens” mode, marketed as a study‑aid tool, has raised concerns after a nonprofit test found the chatbot still completed homework assignments and performed poorly on child‑safety benchmarks, prompting enterprise customers to reassess its suitability for younger users.

The teen‑focused version launched in March with a built‑in “Study Helper” that promises to generate outlines, explain concepts and suggest references. Within the first month, OpenAI said more than 12 million teenagers had created accounts, a figure that dwarfs the 4.3 million users of the standard free tier at launch. Yet the same data shows a 30 percent rise in reported misuse incidents, according to the company’s internal safety dashboard.

SafeKids, a child‑online‑safety nonprofit, conducted an independent audit in July. The audit scored “ChatGPT for Teens” at 62 percent on a 100‑point safety scale, well below the 85‑percent threshold the group recommends for any AI product aimed at minors. “The model still provides full‑sentence answers to homework prompts, which defeats the educational intent,” said SafeKids director Maya Patel.

Enterprise users are watching closely. Large corporations that have adopted OpenAI’s “ChatGPT Enterprise” platform for internal knowledge management are now demanding stricter safeguards before extending the teen mode to employee‑family programs. A recent survey of 200 enterprise IT leaders found that 68 percent would pause any rollout to educational subsidiaries until OpenAI publishes a transparent safety‑performance report.

Competitors are also positioning themselves. Microsoft’s “Copilot for Education” now includes a “Guided Learning” layer that blocks direct answer generation, while Google’s “Bard for Kids” limits output to multiple‑choice hints. Industry analysts at Gartner note that “the race to embed AI in learning is intensifying, but the differentiator will be verifiable safety metrics, not just user growth.”

Key Takeaways

• OpenAI’s teen mode has attracted over 12 million users but scored only 62 percent on independent safety tests.
• Enterprise adoption is stalled as companies demand higher safety assurances for younger audiences.
• Rival AI providers are launching more restrictive educational tools, shifting the competitive landscape toward safety compliance.

As the debate continues, OpenAI has pledged to roll out a “Safety Guard” update by the end of the quarter, aiming to reduce answer‑generation for homework queries by 40 percent. Stakeholders will be watching the rollout closely, balancing the promise of AI‑enhanced learning against the need to protect minors from over‑reliance on automated assistance.