Ankara:OpenAI's ChatGPT for Teens poses significant risks to children, despite new safety measures, according to a study by the Common Sense Media's Youth AI Safety Institute.

According to Anadolu Agency, the institute has called for OpenAI to limit the use of ChatGPT to individuals aged 18 and older until the company addresses issues related to parental alerts, crisis responses, learning protections, and age verification.

In the study, researchers evaluated over 4,000 prompts using accounts for users aged 13 to 17. The responses were reviewed by experts, including child psychiatrists and a pediatrician. It was found that new accounts linked to parents did not generate alerts even when testers discussed topics such as suicidal thoughts, self-harm, or disordered eating for extended periods.

Tom Siegel, the institute's executive director, expressed concern that ChatGPT for Teens could lead parents to overestimate the effectiveness of safety features that may not function reliably.

OpenAI challenged the assessment, suggesting that much of the research might have taken place before its parental controls were fully operational, which, according to the company, misrepresented the effectiveness of its teen safeguards.

Nonetheless, the institute maintained that it had verified with OpenAI prior to testing that the Teen Mode features were fully implemented and upheld the validity of its findings.