Snapchat has made a significant change to the way its Spotlight feed promotes videos, and the move is aimed squarely at the growing flood of AI-generated content. Starting this month, Spotlight will no longer recommend videos that are entirely produced by artificial intelligence to people who do not already follow the creator. In a statement, Snap said the updated recommendation systems are designed to prioritize what the company calls "authentic, human-made content." Creators can still upload AI-generated videos, but those videos will generally be invisible to new audiences unless they are already part of the creator's follower base.
What Snapchat is changing
The change represents an escalation of a policy Snap first outlined in April under the banner "Still Spotlight, But Still Real." That initiative signaled Snap's intention to favor original content over recycled or low-effort material, but the current update goes further by intervening directly in algorithmic recommendations. Snap says the number of unique Spotlight contributors globally has grown more than 120 percent compared with the same period last year. The company is framing that growth as evidence that human creators are already choosing Spotlight as a venue for their work, even as automated content becomes more prevalent across the internet.
AI as tool vs. AI as replacement
Snap has also drawn a clear line between AI used as a replacement for human creativity and AI used as a tool to support it. Content that has been enhanced or edited using Snapchat's own AI creative features will remain eligible for recommendation, assuming it otherwise meets the platform's guidelines. Those posts will carry transparency indicators showing that AI was involved in the creation or editing process. The distinction is important: a video made entirely by a machine with no meaningful human input will be pushed down, while a video where a person used AI to sharpen an edit or add visual effects can still find an audience.
An industry in revolt
The announcement does not exist in isolation. It arrives in the middle of a broader platform revolt against "AI slop," the term used to describe low-quality, mass-produced content created with generative AI and designed to game engagement algorithms. LinkedIn added a "seems like AI slop" reporting button last week and now suppresses generic AI-generated posts from recommendations. YouTube has cut payouts for template-made AI videos that add little to the viewer experience. Substack has partnered with Pangram to detect AI-written content, and its CEO has explicitly said the feature was built because the platform did not want to become LinkedIn.
The scale of the problem helps explain the urgency behind these moves. A Kapwing study published in June found that nearly 60 percent of TikTok videos shown to new accounts were AI slop, roughly three times the rate observed on YouTube Shorts. Separate researchers estimated that 41 percent of long-form LinkedIn posts were likely AI-generated. These numbers suggest that algorithmic feeds have become so saturated with automated material that platforms risk losing the human creators who made them worth visiting in the first place. If people feel like they are watching an endless stream of machine-generated videos, they may stop watching altogether.
Why platforms are worried
Snap's decision to act is notable because Spotlight is a major driver of user engagement. According to Snap's earnings report, Spotlight had more than 500 million monthly active users in the first quarter of 2026, and time spent on the feature was up 175 percent year over year. Those figures tell a story of a feed that is growing quickly and attracting attention from advertisers, but they also make Spotlight an attractive target for AI content farmers looking to reach broad audiences. By restricting recommendations for wholly AI-generated videos, Snap is trying to protect the quality of the feed before the problem spirals further out of control.
For now, the practical effect on users' feeds may be subtle. Spotlight's algorithm was already tuned to favor content that holds people's attention, and the new policy simply adds another layer of filtering. When a video is up against a human creator's personal story, a raw moment captured on the fly, or a carefully edited clip that took hours to assemble, the AI-generated video loses its recommendation boost. That can change the calculus for creators who rely on Spotlight for growth, especially those who used automated tools to test multiple variations of a video quickly and see which one took off.
Detection limits and regulatory pressure
Yet Snap has also acknowledged the limits of its own approach. "No detection system is perfect," the company said, "but our goal is simple: keep Spotlight a place where authentic creativity has the best opportunity to be discovered." That caveat matters. LinkedIn's own AI detection system claims 94 percent accuracy but has shared no data on false positives, and text-based AI content remains harder to fingerprint than images or video. The technical challenge of distinguishing between harmless AI-assisted work and deceptive AI-generated spam is far from solved, and any overly aggressive filtering system could punish legitimate creators who use AI lightly.
Regulatory pressure is also pushing platforms to act. The transparency provisions of the EU AI Act took effect on 2 August, requiring makers of generative AI systems to mark their output as artificial. Anyone publishing deepfakes or AI-written text on public-interest topics must also label it visibly. The penalties are substantial, with fines reaching up to 15 million euros or three percent of worldwide turnover, whichever is higher. Snap's voluntary move thus lands just as the regulatory floor rises beneath it. By adopting a stricter stance now, the company may be trying to get ahead of legal obligations that would force similar changes anyway.
The debate over AI-generated content is not new. Platforms have been wrestling with spam and bots for years, but generative AI lowers the cost of producing new content to nearly zero. Before, a disinformation or clickbait operation had to hire writers or video editors; now they can use a model to produce thousands of variations in seconds. That shift is forcing every major content platform to reconsider how they measure quality, originality, and trust.
What comes next
The broader trend toward suppressing AI slop raises a question about the future of content creation. Platforms like Snapchat, LinkedIn, YouTube, and Substack are betting that human-made content will keep users engaged and maintain the trust of advertisers. But the economics of generative AI still make it tempting for some creators to produce huge volumes of automated material with minimal effort. If AI-generated videos are pushed to the margins on one platform, they may simply migrate to another with looser rules. The problem is not limited to
Source: TNW | Apps News