Automated Content Advisories FAQs

Question 1

What are Automated Content Advisories?

Accepted Answer

Automated Content Advisories are warnings automatically generated for online content to alert users about potentially persuasive or manipulative language. They aim to improve user experience by providing a 'duty to warn' about rhetoric in text, audio, and video content.

Question 2

What types of content can Parabasis generate advisories for?

Accepted Answer

Parabasis can generate advisories for text, audio, and video content. It works with search engines, content platforms, social media, and even AI-generated content from language models.

Question 3

How does Parabasis detect rhetoric in content?

Accepted Answer

Parabasis uses rhetoric models that focus on how things are said rather than specific words. The models err on the side of caution and do not use word embeddings or neural networks for rhetoric detection to avoid biases.

Question 4

What additional features does Parabasis offer beyond rhetoric detection?

Accepted Answer

In addition to rhetoric detection, Parabasis offers theme identification, hate-speech detection, profanity filtering, and nudity detection for text, audio, and video content.

Question 5

How does Parabasis handle user data and privacy?

Accepted Answer

Parabasis stores submitted content for up to 30 days for debugging purposes only. They do not use this data to train ML models or share it outside the company. Only necessary personal information is retained for site functionality.

Question 6

Can Parabasis integrate with different platforms and services?

Accepted Answer

Yes, Parabasis offers a flexible API that can work with various platforms including search engines, content platforms, social media, and AI language models.

Automated Content Advisories Howto

More Information

How to Use Automated Content Advisories