We had a case where the meeting transcript is correct, but the Loom AI recap contains completely unrelated topics, names, and action items that were never discussed in the meeting.
Even more concerning: the names and content in the recap are not related to our company at all.
This raises a serious question about where this content came from, and whether content from our meetings could potentially appear in someone else’s recap.
I already have a high-priority Atlassian support ticket open, but have not received any feedback yet.
Has anyone experienced something similar with Loom AI for Meetings?
Community moderators have prevented the ability to post new answers.
Update
@Martin Runge @Amanda Barber @Arkadiusz Wroblewski
Support is on it. They agreed that the transcript and the recap are totally unrelated.
I have deactivated all Loom AI features for my clients until we know what's going on.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi @Mike Maurer
I have not seen this one. You probably have noted the Loom video ID and have taken a screenshot of the recap before it changes. Have you been recording through the browser plugin?
Can you keep us posted here or by PM, please?
Cheers, Martin
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi @Martin Runge ,
yes, I saved everything:
Additional information:
Yes, I keep you updated where our recaps land ...
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I don't want to be the support engineer that need to figure out if the agent was in the wrong ms teams meeting or else... BUT: for all of us super critical so that we can trust this tool.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hello @Mike Maurer
I’m not sure this is really a data isolation problem based on how this whole thing is set up around Data Isolation between instances. Generative AI can definitely have its Hallucination moments and come up with some wild stuff.
Please address this directly with Support; we are guessing here as we can't see the backend.
Best,
Arek🤠
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi @Arkadiusz Wroblewski ,
as mentioned: support ticket is raised.
The transcript and the recap are totally unrelated. That can't be Hallucinatio.
Cheers, Mike
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Well then, it stays noting; just wait what their investigation says.
Well,lf that recap and transcript are unrelated will rather say it could be a hallucination. You would be surprised how sometimes generative AI makes mistakes.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
I'd recommend raising a support ticket about this to get it in front of the team at Atlassian. Here's the different ways you can get in touch: https://community.atlassian.com/forums/Loom-articles/Need-help-with-your-Loom-account-Here-s-how-to-contact-Support/ba-p/3201201
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
As mentioned above: I've already raised a ticket.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Community moderators have prevented the ability to post new answers.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.