Telling real media from AI-generated media by eye has gotten genuinely difficult. Tools like Sora, Veo, and Midjourney generate images and video convincing enough that the old advice — “look for extra fingers or weird eyes” — barely holds anymore. The tools built to catch this have had to keep up, and while none of them are perfect, a handful of free options are worth knowing about, whether you’re vetting a suspicious image, checking a viral video, or just curious about something that crossed your feed.
A quick caveat before the list: no detection tool is definitive. Treat every result as one input, not a verdict, and cross-check anything important against an independent source.
1. Hive Moderation’s AI Detector
Hive is an established content moderation company, and its free browser-based detector gives quick spot-checks for whether an image or piece of video was likely AI-generated. It’s backed by the same detection models the company licenses to larger platforms, which gives it more credibility than many standalone “free checker” sites built for a single marketing page. It’s best used for a fast first opinion rather than a deep forensic analysis.
2. AI or Not
AI or Not offers a straightforward free image checker: upload or paste a link to an image, and it returns a probability score on whether the image was AI-generated. It’s popular precisely because it’s simple — no account needed for basic checks, and it’s frequently used by journalists and researchers doing quick verification passes on images circulating on social media.
3. Deepware Scanner
Deepware focuses specifically on video, and it’s one of the few free, no-signup options that lets you paste a public video URL directly rather than requiring a download first. It’s geared toward catching face-swap style deepfakes in particular, so it’s a good fit for checking a suspicious video call recording or a clip where a specific person’s face is the concern — less suited to fully AI-generated video where no real footage was involved at all.
4. TrueMedia.org
TrueMedia started as a nonprofit deepfake-detection project and has since been rebooted as a free, open-source tool hosted at Georgetown University. Because it’s nonprofit and academically affiliated rather than a commercial product upselling a paid tier, it’s a solid option when you want a check that isn’t trying to convert you into a paying customer. Video support has been more limited than its image and audio checks, so it’s worth confirming current coverage before relying on it for a specific media type.
5. Content Credentials (C2PA) verification
This one works differently from the other four. Instead of analyzing an image or video for signs it’s fake, Content Credentials is a provenance standard — built by a coalition including Adobe, Google, and Microsoft — that embeds a verifiable history into a file at the moment it’s created, showing whether it came from a camera, was edited, or was generated by AI. Google has been building Content Credentials verification directly into Search and Chrome, and tools built on the C2PA standard let you check an image’s embedded history when that metadata is present. The limitation is real: this only works for content created with participating tools and platforms, and it does nothing for media that never had credentials attached in the first place — which today is still most of what’s online.
A quick reality check on all of this
Detection tools are fighting an uphill battle. They’re generally trained on known generation methods, and newer AI models routinely produce output that slips past classifiers built for older ones — so a “likely real” result isn’t a guarantee, especially for cutting-edge generated content. A few habits matter more than any single tool:
- Use more than one detector when something really matters, since different tools are trained differently and can disagree.
- Check the source, not just the file. Does the same claim or clip appear through an independent, verifiable channel?
- Be cautious what you upload. Free tools vary widely in how long they retain files — check a tool’s stated deletion policy before uploading anything sensitive or containing real people’s faces.
- Don’t rely on visual instinct alone. The “extra finger” era of spotting AI images by eye is largely over for current-generation models.
The bottom line
None of these tools will give you certainty, but together they give you a meaningfully better shot at catching AI-generated content than guessing by eye. For anything with real stakes — a suspicious video call, a viral image used to support a claim, a piece of media tied to a financial request — layering a couple of these checks with independent source verification is the closest thing to a reliable process available right now.
This same “verify before you trust it” instinct is part of the same thinking behind spotting phishing and business email compromise attempts, both covered alongside the broader fundamentals in the Small Business Cybersecurity Risk Checklist.
Leave a Reply