Data Collection Info

Thank you for your interest in Promethean AI! This page contains information about the data being collected by Promethean AI for the purpose of improving your experience with the software and providing you with the level of excellence you deserve. All while fully protecting the privacy and security of your data.

Art Assets

 

No actual art assets are being stored by Promethean AI at any time. No 3d meshes, materials, textures, sounds, animation, concept art, etc… Promethean AI generates a local metadata library that references your local art assets to be used in scene generation. That way you are always in control of your assets and no 3d party has exposure to them.

Semantic Data

 

The only thing we need to store when you train the world-building part of the AI are spatial relationships between objects. I.e. the distance from a table to a chair and their bounds. This data is referred to as Semantic Data and does not include any specific art content but rather the metadata label you assign to it. It is stored on a private AWS bucket that we set up individually for every customer. After processing, the data is used to train the AI which is further shared with your team to ensure the maximum power amplification catered specifically to your team's needs. If deemed general enough by the Promethean AI team parts of the semantic data might be used to train the general model but not before anonymizing all of the labels. No other user of Promethean AI can ever reconstruct your intellectual property.

Usage Statistics

 

Promethean AI posts statistics to a private server on the types of features that are being used and the context of their use cases. This allows us to improve discoverability and efficiency of the features we provide to you. This data is encrypted both in transfer and in storage.

Crash reporting

 

Promethean AI intercepts crashes and sends a call stack up to a third party server(Sentry). This data contains technical Promethean AI information in regards to the crash.

User Logs

 

Promethean AI automatically aggregates user logs on your private AWS bucket dedicated specifically to your data. These can be occasionally accessed for analytics or troubleshooting any issues your team might have.

 

Technical Details

Promethean AI hosts its infrastructure on Amazon Web Services. Our data is always transmitted over HTTPS and encoded in transfer and in storage using AWS AES-256 encryption and multiple additional layers of encryption and data obfuscation to ensure maximum data security.

 

Computer Vision

 

During a computer vision labeling request Promethean AI uploads your asset thumbnails onto our private AWS server to perform the labeling. This thumbnail data can be used to further train our computer vision model to improve your asset management experience. Additionally Promethean AI stores any corrections you make to the labeling system to improve further vision models.

3d Party Services

 

Some of this functionality might be diverted to third party platforms:

  • Voice Recognition

  • Computer Vision (color and shape recognition happens locally)

All CV data that makes its way to 3d party vendors is removed from their system shortly after.

Human Assisted Computer Vision

 

When applying computer vision to assets that our models have low confidence in, Promethean AI may employ a human assisted third party service to help fine tune the labels. This helps us improve your content management experience when working with obscure data. You always have an option to opt out of this when dealing with sensitive data by using the blacklisting option described below.

Blacklisting

 

Promethean AI offers you a blacklist option to ensure that any words contained in the blacklist are never picked up by any of the data collection systems and no images that have the blacklisted words in their metadata ever make it off your local storage.

 

Contact

For any additional questions or concerns don’t hesitate to reach out to info@prometheanai.com