Message to Seller automatically erroneously detected as "profanity".

[lagarto_electronico]

I saw a seller selling some action figure accessories.

I sent the seller a message with a list of what the accessories go to in order to help him sell the items for more money.

For example:
Optimus Prime - Gun
Megatron - Cannon
Soundwave - Missile

This was automatically detected as hate speech and profanity by eBay and I received a message from eBay immediately stating that my account could be suspended!

This message did not tell me what the algorithm detected as profanity and there is no longer any way to reach a human at eBay to find out what is going on!

I want this flag on my account removed immediately! I did nothing wrong!

[southern*sweet*tea]

Maybe the bots thought you were threatening the seller with your megatron?

Artificial intelligence at its finest, folks.

[bigdeals.etc]

---

@southern*sweet*tea wrote:

Maybe the bots thought you were threatening the seller with your megatron?

Artificial intelligence at its finest, folks.

---

Well he IS the biggest terror in the universe.

[lagarto_electronico]

My problem with this is that there is no way for me to know what triggered the action, therefor no way for me to avoid getting flagged again (and potentially suspended). If it is anything like the AI of Facebook, once the AI decides you are an offender, you get put in special category and monitored and eventually anything can be misconstrued.

Additionally, I selected "have an eBay rep call me" twice and they never called me. I used two telephone numbers. No representative ever contacted me.

[bigdeals.etc]

Yea I'd contact ebay specifically. They might not even tell you but it's worth a shot.

On a related topic, sometimes the bots are really stupid. I'll give you an example of how sometimes the bots incorrectly mash up words to form inappropriate words... If you type the sentence "I respect these men", a bot might mash up the end of the word "these" and "men"... and think you typed the word "**bleep**." I've heard of this happening before and the bots find false positive profanity.

Ha ha ha, turns out that word is even prohibited... I can't even show the example. I was referring to the S and E at the end of the word "these" and add it to the last word.

[gwzcomps]

"Super" followed by "very" is inappropriate too  🙄

Also the system recently warned me about sending phone numbers when I tried to put in an eBay item number...

[coolections]

---

@lagarto_electronico wrote:

My problem with this is that there is no way for me to know what triggered the action.

---

How could you not know ? Gun, cannon, and missile all in the same sentence. That should be a tell tale way of knowing. Bots are not always the smartest things in the world.

[tev4all]

 Bots are not always the smartest things in the world.

or, the ones programming the bots.

[bigdeals.etc]

---

@coolections wrote:

---

@lagarto_electronico wrote:

My problem with this is that there is no way for me to know what triggered the action.

---

How could you not know ? Gun, cannon, and missile all in the same sentence. That should be a tell tale way of knowing. Bots are not always the smartest things in the world.

---

I first thought this too seeing the sentence can be mistaken for threatening. But the OP said profanity and I took that at face value thinking those words aren’t profane. But maybe the OP misinterpreted the violation and assumed it only meant profanity when it was probably more general like including hate speech or something.

[burns4thewin]

Bot most likely thought you were threatening the sellers with your gun, cannon & missile.

Can't fix stupid - even when it's a bot.

[shipscript]

@lagarto_electronico 

This is where you have to play a game. Take out all of the spaces and punctuation in your message and then look for new words in that long phrase. That is how the bots work. The words that were detected may have been in the conversational part of your message, not in the list of items. Unfortunately, not all of us have a vocabulary of "hate speech" words to even know what common words have been redefined by the PC crowd.

[lagarto_electronico]

I finally spoke to someone in the Trust and Safety Department in the USA. They reviewed the message and -- even by removing spaces -- could not find what triggered the AI.  The representative stated that the mention of guns, cannons, and missiles did not trigger it.

The only thing I could think of is that I mentioned the Dinobot Sl(a)g. That name is a derogatory term for a loose woman in some ares of Europe. However, the representatives told me they did not think that was it either.

Another possibility is that the bot perceived "Astrotr(a)in" to be a sexual position or "in(s)ecticon" to be a reference to an illegal act with a minor.  Supposedly the issue was flagged and the AI script will be reviewed to see what is creating the false positive.

They did agree that I am not a terrorist. Which is good.

[moondogblues]

I was once flagged for the very same thing and it was because the bots put two words together.  Weird how that works.