Back to today’s briefing

Story perspectives

Experts Urge Disclosure System After GPT-3.5 Flaw Revealed

3/14/2025

50 5

Summary

Story summary

A critical flaw in OpenAI's GPT-3.5 was uncovered in late 2023, revealing issues of word repetition and personal data leaks. In response, over 30 AI experts are advocating for a robust disclosure system to ensure vulnerabilities are reported safely and legally. They caution that ignoring these issues could have dire repercussions for AI technology.

1 of 1

Story summary

A critical flaw in OpenAI's GPT-3.5 was uncovered in late 2023, revealing issues of word repetition and personal data leaks. In response, over 30 AI experts are advocating for a robust disclosure system to ensure vulnerabilities are reported safely and legally. They caution that ignoring these issues could have dire repercussions for AI technology.