ChatGPT Dots texting beta launches without end-to-end encryption

A new mobile channel for dots

OpenAI began rolling out a texting feature for ChatGPT Dots to US Pro users on October 28, 2026. This limited beta allows eligible users to give a dot work through standard phone messaging apps like iMessage or RCS. Instead of creating a new agent for the phone, the system connects the user to an existing dot that already has its own memory and plugins.

The goal of this integration is to provide a convenient way to hand off tasks or add context to work already in progress. A user can ask their dot to keep an eye on something or provide updates directly to their messaging app. Because the channel accepts text, photos, and voice messages, these handoffs can be more complex than simple typed commands.

The dot can also initiate its own updates when it has something relevant to report to the user. It uses the conversation's context to decide where to reach the user, allowing for a more dynamic interaction. This means a user can choose to receive short alerts on their phone while leaving the heavy lifting and long-form project management inside the ChatGPT interface.

How the messaging route works

The technical path of these messages is a critical detail for users concerned about privacy. As Quasa.io reported, a third-party messaging provider called Linq processes the texts on their way to OpenAI. This means the conversation is not end-to-end encrypted between the user and the dot, even if it appears in a familiar phone app.

Because Linq sits in the middle of the data flow, the iMessage or RCS labels on a phone do not guarantee a private connection. Any message sent through this route becomes accessible to the provider during processing. This is a material distinction for anyone handling unpublished transcripts, confidential briefs, or customer details that might be copied into a text message.

The system is designed to be a one-to-one exchange between the account holder and their own dot. It does not support SMS or group threads, which prevents the dot from joining a client team's conversation. Furthermore, the texting channel does not give OpenAI or the dot access to the user's other private phone conversations.

Boundaries for professional use

The rollout has specific limitations regarding who can use the feature and how it interacts with work data. Currently, the beta is excluded from Business and Enterprise workspaces, which is significant for professionals with managed accounts. A personal Pro account does not automatically extend these texting capabilities to a company-managed workspace.

Users should be cautious about where client material crosses the boundary between the phone and the AI. While the dot can access apps and projects already connected to it, sending sensitive material directly into a text thread creates a disclosure risk. For example, a creator could ask a dot to summarize a project it already has permission to see without pasting the contents into the text.

Copying an entire private brief into a text message, however, puts that material directly into the third-party messaging route. The distinction lies in whether the text contains a short instruction or the sensitive material behind the task. A safer approach involves using the text thread for concise task requests while keeping the underlying data within the authorized project connections.

Managing approvals and updates

The texting feature includes a specific workflow for handling actions that require human intervention. If a dot needs approval for a proposed action, it can send a link to the user via text. The user then follows that link to review and approve the request within the ChatGPT interface, rather than doing the work in the messaging app.

This distinction is important because it prevents the dot from taking potentially large actions based on a short text command. The text serves as a notification that a decision is waiting, while the actual review takes place in a more controlled environment. This keeps the messaging thread clear of complex project details while still providing timely alerts.

For online professionals, this setup allows for a streamlined workflow of monitoring and deciding. A user might receive a short alert on their phone when a monitored task needs a decision, then switch to ChatGPT to review the underlying work. This prevents the user from having to manage a long project through a series of message bubbles.