Read.ai sending meeting reports (unsolicited)

According to their website, “Read AI is an AI copilot for wherever you work, making your meetings, emails, and messages more productive with summaries, content discovery, and recommendations.”. I wasn’t familiar with it until recently when a client reached out to advise that they were, all of a sudden, getting Read Message Reports from various Teams, Zoom and Google Meetings that contained sensitive information from those meetings. The from address is executiveassistant@e.read.ai and there doesn’t appear to be any consistency regarding what meetings are impacted. To date, we have confirmed the following:

  • The user does not have the read.ai browser extension / plugin
  • The user does not have the read.ai app installed in Teams, Zoom or anywhere in their Google ecosystem
  • The user does not have an account at read.ai. We tested this by going to the read.ai website and trying to log in but were prompted that the email wasn’t for an existing user.
  • In the most recent impacted Teams meeting, read.ai was not listed as an invitee.
  • The user reached out to read.ai support to detail what was happening and try to find out why and got the following response:

Thanks for reaching out to Read AI. For context, account creation does not happen automatically; based on your feedback it sounds like you created an account to view meeting notes that were shared with you.

When a person using Read invites it to meetings, they have the option to share the meeting report with meeting attendees based on their settings. To ensure that meeting reports are only shared with intended participants, anyone who is not a Read user already will need to create an account to access the meeting notes. Account creation is a multiple-step process (about 13 clicks) that needs to be completed before accessing the report.  Our goal is to be as clear as possible on what is happening in each step of that process.

We want to make it as simple as possible for you to stop or adjust Read:

I am curious to know if anyone reading this uses read.ai (voluntarily / on-purpose or otherwise) or if anyone reading this has found themselves tangled in the same kind of mess that we’re seeing and, if so, if (and how) you were able to unravel it.