seconding this, it failed to correctly identify an AI image which had the DALL-E watermark clearly visible what plans are there to guard against people intenti

seconding this, it failed to correctly identify an AI image which had the DALL-E watermark clearly visible what plans are there to guard against people intentionally poisoning your training data by miscategorizing the images they upload for classification?