Meta, which owns Instagram, announced on its creators’ blog on November 17 that Instagram is rolling out a copyright detection tool for select content creators. The tool scans Instagram and Facebook for copies of a creator’s original videos and lets the creator decide what action to take. This also includes AI-generated content that reuses audio from the original video, as per Meta’s “content protection”. The feature, previously available to eligible Facebook creators, is now expanding to Instagram.

“This space is rapidly evolving and we’re moving quickly on ways to help protect creators. We’re starting to show videos in content protection that match the spoken words in your original video, even when the visuals have been replaced by AI. We’ll begin testing this with a small group of creators in the coming weeks.“ – Instagram

Why it matters: This copyright detection tool might be useful as AI-generated deepfakes and AI slop content feed on the original content as part of engagement farming, a practice of using deceptive or manipulative methods on social media to artificially inflate metrics like likes, comments, shares, retweets, and followers. As a result, a lot of original content creators are also approaching the Indian courts to protect their “personality” from commercial exploitation as well as from AI deepfakes. 

However, how effectively this tool can help original creators against deepfakes or AI slop depends on whether other content creators intend to reuse the content. For instance, reaction videos could be one of the genuine use cases where the distinction between original and copied content can be difficult to establish, which could disproportionately impact the content creators.

‘AI stealing original content’: what does it mean? Basically, AI-generated copies of original content, including deepfakes and low-quality AI-generated material. For instance, in October, Andrew Scott, an interdisciplinary artist and content creator, wrote on Meta’s Threads that AI deepfake replica videos of his original work “got 50 million views and 1 million likes in just a few days.” He also posted the video of AI-generated content compared to the original video.

Responding to this, Adam Mosseri, the Head of Instagram at Meta, who oversees product, engineering, and operations for the platform, said: 

“To be clear, I am very worried about this trend. We are actively working out what to do, it’s just a tricky problem.

We already let people make copyright claims against content, and take stolen copyrighted material.

That’s not enough though, as I assume many comedians and dancers haven’t gone through the trouble to copyright their material. We need a way to help those creatives without letting bad actors make more fake claims (happens all the time unfortunately).” – Adam Mosseri, Instagram Head.

How does it work?

Instagram says content protection scans both apps for visual matches of a creator’s protected videos. The tool targets copies that reuse something from a creator’s own video. Instagram’s test extends that from visuals to spoken words, so a copy with AI-generated visuals but the same speech can still surface. It uses the same matching technology as Meta’s Rights Manager, a service for copyright holders. 

Who is eligible?

However, not every creator has access to this tool. YouTube’s Content ID system also has eligibility restrictions, such as the Content ID system. Nonetheless, once a creator enrols in the programme, every original video they post after the enrolment is automatically protected.

“Content protection is available to Instagram creators who satisfy our eligibility criteria, which is largely based on meeting our originality and integrity standards.” – Instagram. 

Alternatively, creators can also add older videos manually. When the tool finds a match, it will show how many views the copy has and how many followers the posting profile has.

For the Facebook version, Meta said in 2025 that creators who file false reports could face restrictions on their accounts or lose access to the tool. Reports on the Instagram launch do not say whether the same penalty applies. They also do not describe how the account whose video is hidden can contest the decision.

Instagram’s foray into long-form content: 

On October 26, Instagram head Adam Mosseri and Tessa Lyons, Instagram’s vice-president of product, announced in a video that they are testing long-form videos on Instagram called Features to see ” how they fit with Reels, Series, Live and TV.” 

Instagram’s short-form Reels can be exceptionally good at grabbing attention, but they might attract only short-lived audiences. Consistent long-form content can primarily build trust, habit, repeat viewership, and monetizable loyalty over time, because it draws people to the content regardless of algorithmic push.

YouTube has developed a substantial creator economy around long-form videos. Instagram’s additional copyright protections against potential infringement and AI deepfake exploitation can also incentivise people to create an audience for long-form content on Instagram. 

 How does it compare with YouTube’s Content ID system?

YouTube has its own Content ID, an automated system for copyright owners. YouTube says it scans every upload against a database of audio and visual files submitted by owners. A match gets a claim that blocks the video, monetises it with ads, or tracks its viewership, depending on the owner’s settings. Original content creators or owners must meet criteria that include holding “exclusive rights to a substantial body of original material” that is frequently uploaded to YouTube. 

On the scale of Content ID usage, YouTube said over 99% of Content ID claims in 2024 came from automated detection, and owners chose to monetise over 90% of claims, as per a Digital Music News report. 

Alongside the traditional Content ID system, YouTube handles facial data protection through its likeness detection tool, which works similarly to Content ID, YouTube’s help page says, but searches for a creator’s likeness instead of copyrighted content. However, to be eligible for this, content creators should verify with a government-issued ID and a short selfie video, which are used to verify identity and support likeness matching.

Similarly, YouTube launched the tool for Partner Programme creators in October 2025, MediaNama reported. It extended it to journalists, officials and candidates in March, and to the entertainment industry in April, as per the TechCrunch report. In May, YouTube said it would open the tool to everyone over 18, Dataconomy reported. At its Made on YouTube event in September, it said speaking-voice detection would follow later in 2026.

Writing about the gaps in YouTube likeness detection tech, MediaNama’s editor and founder, Nikhil Pahwa, wrote: 

“What if AI likeness is used in parody? For example, if I make a parody review of a Maruti 800 (an Indian car no longer manufactured by Maruti Suzuki) using an AI likeness of MKBHD, he can’t take it down under copyright law because it’s a parody, but he can get it taken down because of likeness. YouTube, especially in India, has largely failed to protect fair usage. According to YouTube’s policies, if the video is real, a copyright claim must be made (taking into account fair use and other copyright exceptions), but if it is AI-generated, one can make a privacy claim. Does this mean that human mimicry is kosher, but AI generated mimicry is not? Why the discrimination? While this arms creators and copyright owners with an additional tool, it harms fair usage and adds identity-based gatekeeping.” – Nikhil Pahwa.

Also read

The post Instagram launches copyright detection tool for AI deepfakes and stolen content appeared first on MEDIANAMA.