Apparent contradiction in guidance on training a proprietary never released model on -ND content

## Description

According to the -ND section of [this document](https://creativecommons.org/using-cc-licensed-works-for-ai-training-2/) training on -ND data (in so far copyright applies) is prohibited:

> [CC BY-ND](https://creativecommons.org/licenses/by-nd/4.0/) and [CC BY-NC-ND](https://creativecommons.org/licenses/by-nc-nd/4.0/) prohibit creating derivative works. Following the NoDerivatives restriction would require that ND-licensed content not be used as training data.

However at the end of the document it reads:

> Step 2 – When Do CC License Conditions Apply?
> - The attribution (BY) and ShareAlike (SA) conditions, and NoDerivatives (ND) restriction are triggered only when works or adaptations of works are publicly shared.

This looks to be a direct contradiction to me. The later being able to be read as "training proprietary model on ND content is fine as long as you do not get caught". I am not sure this interpretation should be left unchallenged.

## Expectation
I would like at least one these two sections changed so they no longer allow a contradictory reading.

## Resolution
I am not a lawyer and in no way able to judge what change is appropriate
- [ ] I would be interested in resolving this bug.