OpenAI has announced a series of enhanced safety measures designed for under 18’s using ChatGPT to introduce controls that limit how “human” the chatbot can appear amid growing concerns about the use of AI among children.
Those with the new teen account can disable voice responses to ensure the tool is understood as AI rather than a person. OpenAI is also introducing reminders for young people to take a break when using the tool to prompt healthier usage patterns.
Reinforcing real-world relationships
OpenAI has said the update was not triggered by any single incident but the update comes following a wider concern about people believing AI systems to be sentient with the company reinforcing protections that should ‘reflect developmental stage, reinforce real-world relationships and support healthier use over time.’
Teen accounts will become a default for anyone who identifies as under 18 during sign up, as well as for users OpenAI determines to be minors. Children under 13 remain unable to use the product.
Heather Barnhart, Senior Digital Forensic Expert at Cellebrite and SANS DFIR Curriculum lead commented: “Technology policy makers are up against what could be the greatest generation in tech. AI adoption is being introduced across all sectors and the children who understand how AI works and how to use it responsibly will be at an advantage as AI increasingly becomes a life skill. With this in mind, regulation should prioritise educating children on how to use it safely and what potential red flags to look out for.”
Providing direct answers
“Parents, educators and social media companies need to take an active role in guiding children on use. This gets at the larger issue, which is parents having open conversations about the dangers that lurk online - and having concrete guardrails on what their kids are doing and how they’re interacting with technology that is having an active role in shaping every sector. With education and healthy engagement, families can help their children navigate the online world, responsibly and safely.”
As part of this, OpenAI is also expanding its Study Mode which was launched last year to help students work through problems without providing direct answers and set specific hours where Study Mode will automatically activate. OpenAI has reported nine out of 10 of its younger users use it to help them with learning.