I am playing with gpt4v for feeding a screen shot and ask what it can inference, then receive {‘error’: {‘message’: ‘Your input image may contain content that is not allowed by our safety system.’ ‘code’: ‘content_policy_violation’}}. The screen shot is from powerpoint. [cropped_1090_395_screenshot…

It looks very CAPTCHA-like, as if it could be a bot test. It seems now OpenAI is using a vision moderator to simply block suspect images from being received, instead of the AI model trained to deny half of the images you send.

Hi and welcome to the Forum! I took the liberty to run a GPT-4 Vision request locally with the above screenshot. By simply using the prompt “Please provide a description of the contents of this picture” I faced no issue and the following was returned. Perhaps the way you phrased your message cause…

Thanks, that really resembles what you have said.

Policy violation while playing with GPT api

API

waterdeeper2 March 23, 2024, 2:08pm 5

Thanks, might be. Next time i will try your prompt.

Topic		Replies	Views
'Your input image may contain content that is not allowed by our safety system.' -Vision API response Bugs gpt-4 , gpt-4-vision	15	4640	March 11, 2024
GPT4V and content_policy_violation API gpt-4-vision	4	3869	February 8, 2024
Vision API - "Image not allowed by our safety system" Bugs gpt-4-vision	9	1740	May 29, 2024
GPT-4v returns "I'm sorry, I can't provide help with that request." for one photo API gpt-4-vision	3	1344	March 12, 2024
ResponsibleAIPolicyViolation Error due to particular prompt and image passed Community gpt-4 , gpt-4-vision , api-vision	1	1679	September 27, 2024

Policy violation while playing with GPT api

Related topics