# OpenAI Sets Early 2027 Target for First Community Safety Report

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/45000
Published: 2026-10-30T18:30:23.000Z
Updated: 2026-10-30T18:30:23.000Z
Section: Technology

> The series will examine violence prevention and the effectiveness of OpenAI’s safeguards, with new installments planned every six months.

OpenAI plans to begin a recurring public report on community safety in early 2027. The first edition will cover the second half of 2026, with later installments planned every six months.

The series will examine violence prevention, broader community-safety issues and the effectiveness of OpenAI’s safeguards. The company plans to explain how its systems detect safety concerns, investigate them and respond. Reports are also expected to include anonymized examples, data, mitigation measures and policy changes.

The planned coverage combines examples of safety issues with information about the measures OpenAI uses and changes to its policies. The company says the reports will assess its safeguards alongside its violence-prevention work, giving the series a recurring focus on how it addresses community-safety concerns.

OpenAI also disclosed that it made more than 200 referrals to law enforcement in the third quarter of 2026 involving imminent threats of violence. The announcement provided no further breakdown of the referrals.

The announcement also outlined tighter limits on ChatGPT responses about past mass attacks. The changes include restrictions on weapon details and information about obtaining weapons. They apply regardless of a user’s age or reason for asking.

The early 2027 edition will be the first in OpenAI’s community-safety report series, not the company’s first report. OpenAI published OpenAI Signals in February 2026.