Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
ChatGPT “adult” filters are buggy
Sources told the WSJ they doubted OpenAI’s tools were ready to exclude children from banned content.
Your whistleblowing comes later OpenAI fired a senior security executive who opposed the launch of “adult mode”. OpenAI denied that the firing was related, but the outgoing employee directly criticized both the AI company’s ability to block children from content and prevent products from promoting child exploitation. Additionally, a second former security staff member also spoke last fall, warning that parents should not trust OpenAI’s “adult mode” claims.
To counter this narrative, OpenAI’s spokesperson promised that the company “has developed a plan to monitor a variety of potential long-term effects of adult mode, both positive and negative.”
However, that plan was likely developed with the same experts who the WSJ reported strongly oppose its implementation, leaving parents wondering whether OpenAI cares about the advice of its youth welfare team or not.
In addition to ineffective age controls or smart minors circumventing age barriers, OpenAI may run into trouble with parents if its own systems fail unexpectedly. In April, when OpenAI began dabbling with more daring results, OpenAI fixed a bug that TechCrunch Tests found was allowing minors to access graphic erotic content on ChatGPT. It appears that OpenAI filters that were supposed to clearly restrict “sensitive content like erotica to limited contexts like scientific, historical, or news reporting” were broken.
“In this case, a bug allowed responses outside of those guidelines, and we are actively implementing a fix to limit these generations,” OpenAI said at the time.
OpenAI did not respond to Ars’ request for comment.
If you or someone you know is suicidal or distressed, call the Suicide Prevention Lifeline number by dialing 988, which will connect you to a local crisis center.
