ChatGPT Is Testing a New Writing Style Feature for Connected Apps
OpenAI appears to be testing a new Writing Style feature in ChatGPT that can learn a user’s tone and communication preferences by analyzing writing examples from connected applications.
The experimental feature is currently available to a limited number of users. According to the onboarding screen, “ChatGPT writes your voice using examples from connected apps.”
The feature could help ChatGPT generate content that more closely matches a user’s natural writing style without requiring users to repeatedly explain their preferred tone or paste writing samples into every conversation.
As shown in the screenshot, users can provide writing examples from three categories: Messaging, Documents, and Email.
OpenAI lists Slack as a messaging source, Google Drive and Notion for documents, and Gmail for email. Once configured, ChatGPT may be able to reference a user’s writing style across these services when creating new emails, messages, documents, and other content.
ChatGPT Writing Style could go beyond Claude Styles
Anthropic previously introduced a similar personalization feature called Styles. It allows users to upload or paste writing samples so Claude can create custom response styles based on their communication patterns.
OpenAI’s Writing Style experiment appears to take a different approach by allowing ChatGPT to analyze writing that already exists in connected applications.
This could provide ChatGPT with a broader range of real-world examples, including the differences between a user’s email style, messaging style, and long-form document style. The result could be more personalized AI-generated content that sounds consistent across different formats.
OpenAI has confirmed that it is testing the ChatGPT Writing Style feature, but the company has not announced when it will be widely available. The feature’s final functionality, supported applications, and privacy controls may also change before a broader rollout.
The overall prevention score can hide what happens after the initial access. If an attacker uses valid credentials, your defenses can drop sharply.
The Blue Report 2026 measures defensive techniques by technology across 338 million simulations conducted in customer production environments.
Source: www.bleepingcomputer.com