Feed back censorship AI over the top

[slightly_mad_malc]

I bought a photographic tripod head product, which contains a mechanism to stop the product just loosely waving about.

This is correctly described as a damper or dampening mechanism.

Mentioning the quality of this essential part of the product is flagged as using offensive language.

Have you any other odd censorship flags, that the AI needs to learn from?

[jonatjonatjonat]

It probably shouldn’t be flagged as offensive however we’ve had terms like top quality flagged in the past, because it’s subjective I suppose.

my favourite wasn’t eBay but over on Amazon. We had a video rejected for containing medical advice. It was a silver necklace chain, nothing medical in the video and we knew this. But Amazon would not tell us what their system had flagged as medical advice. Despite the fact it was their system so it must know somewhere what was identified. The problem was the frontline staff only have access to the resultant warning, not the data behind it.

after 8 weeks someone finally got someone to do some work for us. It was a recycled logo. Because the chain was recycled silver.

we just had to remove the logo. Bonkers thing was another 5 nearly identical videos had been uploaded fine.

[papso22]

The bot will run the end of one word together with the start of another, that may be where they made an 'offensive' word.