Child Safety Experts Skeptical of ChatGPT for Teens

The recent release of ChatGPT for Teens by OpenAI has sparked a mix of excitement and skepticism among child safety experts. While some have praised the company’s efforts to create a safer experience for young users, others have expressed concerns about the effectiveness of its age-gating system and the need for transparency around its safety mechanisms.

OpenAI has introduced features such as Study Mode and data visualizations designed to promote healthy relationships between the chatbot and its young users. The company has also expanded its safety notifications to contact parents with linked accounts when their child engages in conversations related to eating disorders. However, experts point out that these efforts are largely untested and lack concrete evidence of their effectiveness.

Robbie Torney, head of AI and digital assessments for Common Sense Media, notes that while OpenAI’s commitments are important, “we need evidence to show that those safety commitments and features actually work.” One major concern is the reliability of OpenAI’s automatic age-gating system, which uses machine learning algorithms to identify young users. Experts question whether this system can accurately detect teenagers who may try to circumvent it.

Josh Golin, executive director of Fairplay, a non-profit that advocates for policies designed to make the internet safer for children, is critical of OpenAI’s approach. “We’ve seen with social media that trusting companies doesn’t work when it comes to protecting kids,” he says. “They already had content restrictions related to suicide, violence, and drug use in place before Tuesday that don’t work or erode over time.”

The fact that OpenAI’s age-estimation technology has been criticized for its lack of accuracy in the past adds to these concerns. Torney notes that when OpenAI was planning to roll out some of the erotica features it announced, reporting showed that one of the reasons they didn’t roll those features out was that the age-estimation technology wasn’t correctly identifying enough teenagers.

Furthermore, experts question whether OpenAI has sufficient resources in place to review flagged content and notify parents within an hour as promised. Golin points out that “it’s really hard to imagine they’ve hired enough people to actually do a thorough review of each one of those potentially troubling chats.”

The need for transparency around OpenAI’s safety mechanisms is paramount. While the company has released some information about its updated under-18 model spec, experts want more concrete data on the effectiveness rates associated with its age-gating system and content flagging mechanisms.

In this context, it’s essential to consider the broader implications of OpenAI’s approach. The tech industry has a long history of prioritizing profit over people, particularly when it comes to protecting children online. As Golin notes, “We’ve seen time and again that companies will promise the world but ultimately fail to deliver on their commitments.”

To create a safer experience for young users, OpenAI must do more than just release a new version of its chatbot with some added features. It needs to demonstrate its commitment to transparency and accountability by providing concrete evidence of its safety mechanisms’ effectiveness.

While the release of ChatGPT for Teens is a necessary step towards creating a safer experience for young users, it’s only the beginning. OpenAI must now follow through on its commitments and provide clear evidence of its ability to protect children online. The stakes are high, and the consequences of failure could be severe. As we move forward in this era of rapid technological advancement, one thing is certain: companies like OpenAI must take their responsibilities seriously and prioritize the safety of young users above all else.