Meta’s new AI detection tool doesn’t quite work as advertised, according to a new report.
Meta introduced its first imaging model, Muse Image, earlier this week. As part of the debutThe tech giant also announced that all images generated by the model would include an invisible watermark system called Content Seal. This signal would remain intact even when users “crop, compress, resize or screenshot” the AI-generated image, the company stated. To help pick up the Content Seal signal, Meta also announced that it was previewing an AI detection tool to check if an image was generated by Muse Image.
But, in a report released Friday, Reuters Journalists found that the AI detection tool failed to identify more than half of the images it generated once cropped. In the first test, Reuters found that the tool correctly identified all 40 images generated by Muse Image as AI-generated, but once those images were cropped to half or a third of their original size, the tool was only able to identify 55% as AI-generated.
As generative AI tools get better at producing jaw-dropping deepfakes, detection becomes more a more complicated problem to solve. According to cybersecurity company deep blowThe volume of AI-generated online deepfakes has seen annual growth of approximately 900% between 2023 and 2025. But detection capabilities have not fully advanced in parallel with this surge in popularity. Commercial AI-powered detection tools are still plagued by errors, while the average person’s ability to identify AI-generated content is no better than flipping a coin, according to previous studies.
Although it’s not a true immediate success, that’s the gap Meta aims to address with the new Content Seal and its detection tool.
Muse Image and its accompanying products were intended to be a big step forward for Meta, which has arguably been behind its competitors in the AI space. Last year, Meta CEO Mark Zuckerberg decided there was no time like the present to try to catch up and announced a major AI recovery effort. The turnaround plan included committing multibillion-dollar investments in research and development and poaching top talent from rivals across the industry, all in the name of creating better AI products and the noble goal of creating artificial superintelligence.
An additional couple of billion dedicated to AI and some more restructurings Meta later unveiled the first major fruit of that work in April with spark musea proprietary model that it said it plans to open source in the future and that had a mixed reception. Another big one was this week’s Muse Image debut.
But the debut of the imager and its accompanying tools has been mired in controversy, and not just because of the Reuters report. Instagram users were alarmed when they discovered that the AI model could use photos from any public profile without explicitly asking the owner of said profile for consent. that feature has now been removed.
The company is now considering its next big debut in generative AI: a video generator called Muse Video. We hope the company can close any gaps in detection tools and adequately address user privacy concerns before that model goes away.
