Copyleaks, which already offers plagiarism and AI-text detection tools, is rolling out a new system to detect whether images have been generated or manipulated by AI. It gives users a probability score for how much of an image is AI-altered and highlights affected areas. Fraud risks like fake receipts or altered documents are driving this development. The company says the tool is relevant for sectors such as education, finance and publishing. That said, early tests showed mixed results: some AI-generated images were flagged correctly, but the system also produced false positives and missed some altered images.
