• Source:JND

The company has announced a major update to its Model Spec, the internal rulebook that governs how its AI responds in real-world conversations. With the latest revision, safety for users aged 13 to 17 now takes absolute priority, even if that means dialling back other long-standing goals like maximum helpfulness or open-ended engagement.

The update comes as conversations around AI exposure for minors continue to intensify worldwide, with regulators, parents, and educators all questioning how safe these tools really are for younger users.

ALSO READ: YouTube Bans Indian Channel Linked To Fake AI Generated Movie Trailers: Here's What Happened

What The Model Spec Actually Does

While it sounds technical, the Model Spec plays a huge role in how ChatGPT behaves. It’s a publicly available document that lays out OpenAI’s values, boundaries, and behavioural expectations for its AI systems. When ChatGPT decides how to respond to sensitive questions, emotional topics, or high-stakes prompts, this document is often the deciding factor.

OpenAI says the update was overdue. Teenagers, the company notes, don’t interact with technology the same way adults do, and conversations involving minors need extra guardrails. “While the principles of the Model Spec continue to apply to both adult and teen users, this update clarifies how it should be applied in teen contexts, especially where safety considerations for minors may be more pronounced,” the company said.

Four New Rules For Teen Conversations

At the heart of the update are four new rules that kick in whenever ChatGPT identifies a teenage user.

Teen safety comes first, even if it conflicts with other goals.

ChatGPT should encourage real-world support, nudging teens toward trusted people and offline help when needed.

Teens should be treated like teens, not talked down to, but not treated as adults either.

The AI should be transparent, setting clear expectations about what it can and cannot do.

The first rule is the most significant shift. OpenAI has made it clear that when teens are involved, safety now overrides its traditional focus on being as helpful or flexible as possible.

Less Agreement, Clearer Limits

Another subtle but important change affects how agreeable ChatGPT behaves with teenagers. OpenAI says it is reducing the system’s “sycophancy” for users aged 13 to 17. In simple terms, the AI will be less likely to blindly agree with or validate teen opinions if doing so could lead to harm.

This is aimed at preventing situations where the chatbot unintentionally reinforces risky ideas, emotional spirals, or poor decisions.

Age Detection Is Expanding

Alongside the rule changes, OpenAI confirmed that its age prediction model is now in early rollout. The system looks at conversational cues to estimate a user’s age when it hasn’t been explicitly shared. The company says this will soon be active across ChatGPT’s consumer plans.

ALSO READ: Instagram Update: Hashtags Limited On Reels And Posts! This Is What It Means For You

Anthropic is moving in a similar direction. While Claude is currently limited to users aged 18 and above, the company has acknowledged it is working on its own underage detection tools.

Taken together, these moves point to a wider shift across the AI industry. As chatbots become more deeply woven into everyday life, companies are under growing pressure to prove that speed, scale, and intelligence don’t come at the cost of safety, especially for younger users.


Also In News