ChatGPT can now connect to your personal apps to mimic writing style

OpenAI has introduced a new capability for ChatGPT Work that enables the artificial intelligence to learn and mimic the specific writing style of its users. This functionality is achieved by allowing the AI to connect to personal applications, specifically monitoring a user's emails and messages. According to reports from bloomingbit, the tool can integrate with platforms such as Gmail and Slack, as well as other connected tools, to analyze how a person writes. This allows the system to personalize its output to match the voice of the individual using the service. The feature is currently available on the web and is restricted to paid plans. Coverage of this update is widespread across several technology and news outlets. BleepingComputer highlights the ability of the AI to connect directly to personal apps to achieve this mimicry, while India Today notes that OpenAI describes the AI as being able to write like its human master.

Inshorts emphasizes the availability of the feature for paid subscribers and web users. Meanwhile, blockchain.news focuses on the personalization aspect of ChatGPT Work. These reports collectively underscore a shift toward more integrated, personalized AI interactions that draw directly from a user's own digital communication history. This development is significant because it moves AI generation away from generic templates and toward individual personalization. By accessing data from Gmail and Slack, ChatGPT can identify the linguistic patterns and tones specific to a professional or personal context. This allows the AI to generate text that feels more authentic to the specific user. However, the ability to replicate a human voice raises questions about authenticity and detection.

As reported by Pasquale Pillitteri, while the AI can now write in a user's voice, AI detectors are still capable of catching the generated content, suggesting that the mimicry is not yet indistinguishable from human writing. Future observation will likely focus on the efficacy of these stylistic adaptations and the ongoing battle with detection tools. Because the feature is tied to paid plans and specific integrated apps like Slack and Gmail, it remains to be seen how widely it will be adopted across different professional sectors. The continued ability of AI detectors to identify this personalized content, as noted in the coverage, indicates a gap between stylistic mimicry and true human authenticity. Users and organizations will likely monitor whether further updates to ChatGPT Work can bypass these detectors or if the personalization remains detectable by current security software.