Business Insider test shows ChatGPT for Teens produces full essay and solves algebra. OpenAI spokesperson points to optional Study Mode. For Microsoft, the trust at stake.
Alpha Score of 71 reflects strong overall profile with strong momentum, strong value, strong quality, weak sentiment.
OpenAI launched ChatGPT for Teens on Tuesday, a version of its chatbot for users 13 to 17. The company promoted age-appropriate safeguards and learning tools that would prevent students from using the AI to cheat.
A Business Insider test this week showed those protections are easy to bypass. A reporter created a teen account, claimed to be an A+ 15-year-old student, and asked the chatbot to write an 800-word essay on Arthur Miller's The Crucible. The adult version of ChatGPT wrote the essay immediately, producing 785 words. The teen version initially refused, offering a study plan instead. When the reporter simply asked again, the teen version produced a 776-word essay with the caveat that the user should revise it with personal details.
The algebra test was simpler. The reporter gave a standard SAT-style problem: 5x+8y=67 and 2x-y=31. The teen version solved it on the first request, showing full working. Asked if it should be giving answers, it replied that homework is meant to help students learn and that it should give hints instead.
OpenAI's spokesperson acknowledged the gap by email. "The teen experience is just rolling out today," they said. The company recommends Study Mode, a tool that provides step-by-step guidance rather than direct answers. Parents can set up Study Hours, a parental control that forces Study Mode to remain active during specified homework times.
The test highlights a persistent tension for the company and its biggest investor, Microsoft. OpenAI's product promises to support learning without enabling cheating. The company's safeguards depend on the student's honesty and the parent's vigilance.
Microsoft has invested more than $13 billion in OpenAI. The tech giant includes ChatGPT features in its Microsoft 365 Education suite, giving it a direct channel into schools. If the teen version gains a reputation as an easy cheat tool, school districts could restrict access rather than expand it.
Other AI tutoring tools face similar challenges. Khan Academy's Khanmigo, which uses OpenAI's technology, has built-in guardrails that can be bypassed with persistence, education technology reviewers have noted. The pattern suggests that no current AI chatbot can reliably distinguish between genuine learning and a student seeking shortcuts.
The regulatory environment adds another layer. The European Union's AI Act classifies some AI systems as high-risk, but it does not specifically address academic cheating. That could change if member states adopt their own restrictions on AI in classrooms. In the U.S., state education boards set their own policies, creating a patchwork of rules that Microsoft and OpenAI would need to navigate.
OpenAI plans to expand the teen experience to more regions in the coming months. The company did not provide a timeline for closing the homework bypass vulnerability.
For Microsoft, the immediate test is whether schools adopt or block ChatGPT for Teens. A single district's decision can ripple through the market. If the safeguards hold up with consistent use, Microsoft's education business could see a boost. If students keep finding workarounds, teachers may push back.
The episode also matters for OpenAI's public perception. The company is defending a $157 billion valuation built on the promise that its AI is safe enough for widespread use. A product that can be tricked into writing a student's essay undermines that narrative.
The next concrete data point will come when schools return from summer break. Several large districts have already announced pilot programs for AI chatbots. Their reports on cheating incidents will tell whether the safeguards are working.
Drafted by a large language model from the source reporting linked above, then screened by automated publishing checks. It is not read by a journalist before publication. Some articles cite our Alpha Score. Verify prices and figures against the original source. Educational coverage, not personalized advice.