AI images often have small imperfections, which you can see if you where to look. AI text, however, is pretty difficult to detect. Most tools that try to detect AI-written content are about 50% accurate at best (so: a coin flip).
Of course, Midjourney, OpenAI and other companies will always have a record of what their systems generated. They might be the best ones to verify AI-generated content!
Tags will need to be created in order to detect AI, the only problem is who should create it?! That’s a lot of power to have and I don’t trust the company that creates it and certainly don’t want the tags created by the government. The big question will be how we navigate that security issue.
Does AI generated data/images leave a trail?
AI images often have small imperfections, which you can see if you where to look. AI text, however, is pretty difficult to detect. Most tools that try to detect AI-written content are about 50% accurate at best (so: a coin flip).
Of course, Midjourney, OpenAI and other companies will always have a record of what their systems generated. They might be the best ones to verify AI-generated content!
Tags will need to be created in order to detect AI, the only problem is who should create it?! That’s a lot of power to have and I don’t trust the company that creates it and certainly don’t want the tags created by the government. The big question will be how we navigate that security issue.