OpenAI’s Parental Controls Put Up Guardrails for Teens’ Chatbot Use

Published 15 October 2025

Author
Nienke Creemers
2 min read

Amidst rising concerns about youngsters’ exposure to violent or sexual content from artificial intelligence (AI) chatbots, OpenAI (US) has introduced parental controls for ChatGPT. With 72% of American teens and 64% of nine- to 17-year-old Brits regularly engaging with chatbots, the update seeks to protect prolific young users (Common Sense Media & Internet Matters, 2025).

  • New Teen Safety Features: OpenAI’s new parental controls allow parents to link their account to their teen’s (age 13-17) and customise the safety settings. Once activated, minors can’t encounter graphic, sexual/romantic, violent and extreme beauty-related content. Parents can also disable Voice Mode and Image Generation, set quiet hours (times when the account can’t be accessed) and prevent their child’s chats from being used to train ChatGPT. Notably, if ChatGPT detects signs of self-harm or violent ideation, parents are alerted via email, text and/or push notifications. Despite the potential utility of these parental controls, it’s worth noting that teens can easily avoid them by creating a new OpenAI account.

  • Teens’ Chatbot Usage: OpenAI’s introduction of child protection features come after a 16-year-old American committed suicide after “months of encouragement from ChatGPT”. The teen’s family has since filed a lawsuit against OpenAI.

    The case highlights broader concerns – not just about what teens might see, but why they’re using AI. Young people who turn to chatbots for companionship, advice and emotional connection may be especially vulnerable to harmful content. Among British nine- to 17-year-old chatbot users, 23% ask AI for advice, 18% simply chat, 6% want a friend, and 3% seek emotional support (Internet Matters, 2025).

  • Parents Are Worried & Acting: Concerns about AI usage are mounting among parents. In the UK, 51% of parents fear that chatbots expose their kids to dangerous ideas (like self-harm, suicide or drug-taking), and 49% worry about sexual conversations (Internet Matters, 2025). Fifty-one per cent have discussed how AI chatbots aren’t real people with their children, and 31% have addressed potential wellbeing risks (Internet Matters, 2025).

 

See our Future Consumer Life Cycle 26/27: Gen Alpha for more on kids and their AI usage, and Exploring Modern Parenthood for further insights on parents’ attitudes towards AI.