Scrolling at night, you come across a short video about “the latest rules on personal finance.” The title is urgent, the visuals are polished, the narration is smooth, and a label underneath says “AI-generated.” When it ends, you can’t say which rule it cited, who said it, or whether to believe it. The next video, and the one after that, are more of the same.
Platforms add AI labels so viewers are informed. But being informed and saving your attention are different things. A label tells you “this may have been made with AI”; it doesn’t decide whether it’s worth watching. What actually cuts down low-quality content is changing the question. Instead of “was this made by AI?”, ask “does it have sources, is someone accountable for it, does it state its limits?” Then use that standard to reshape where your information comes from.
A label says how it was made, not whether to trust it
Major platforms really are labeling. Since 2024, YouTube has required creators to disclose altered or synthetic content that could easily be mistaken for real, such as swapped faces, synthesized voices, or fabricated footage of real events. The label usually appears in the video description, while sensitive topics like health, news, elections, and finance get a label on the video itself. Meta adds an “AI info” label when it detects industry-standard AI image indicators or when users disclose it themselves.
These labels are valuable, but they answer “how was this made,” not “can I trust it.” A video labeled as AI may be clearly organized with complete sources; an unlabeled video may be filmed by a real person and still be entirely misleading. And once something is labeled, the same kind of content keeps getting recommended to you.
TikTok went a step further in November 2025, adding an AI content slider to its “Manage Topics” settings so users can adjust how much AI content appears in their For You feed. But according to Social Media Today, the setting can only reduce AI content, not switch it off, and AI content the system fails to detect still appears. In other words, what platforms offer is still a volume knob, not an off switch.
What to block is content with no line of accountability
When judging an item, a more useful question than whether AI made it is: if I believe this, will I make a worse decision? Check it with three questions:
- Does it have sources? Can what it says be traced back to the original rule, study, official announcement, or interview?
- Is someone accountable for the judgment? Can you tell who put it together and what they based their choices on? Or is it just several articles squeezed into a summary that looks complete?
- Does it state its limits? Does it say where it doesn’t apply, or what information may be out of date?
Content that fails all three, whether or not AI made it, doesn’t belong in your daily information diet. Conversely, content whose author explains that AI was used only for captions or translation, with complete sources and methods, is still worth reading even with an AI label.
The same three questions apply to what you produce. If you use AI to draft or translate, readers will check you by the same standard: where are your sources, where is your judgment, where are your limits?
Use three questions to reshape your entry points
Platform filters are still incomplete, but anyone can adjust their own entry points now by sorting content into three groups:
- Avoid: content with no sources, that draws clicks through fear or exaggerated titles, or that looks like a tutorial but has no steps and no named author. Unfollow accounts that only repost summaries on social platforms, and tap “Not interested” on such videos.
- Skip: nothing wrong, but nothing new either, just the news repackaged. No need to be annoyed; just don’t spend your deep-reading time on it.
- Keep: authors, organizations, and long-term observers who clearly state their sources, judgment, and limits. Put them in a list or subscribe.
For important questions, such as decisions about health, money, or the law, go to official documents, research, or long-trusted media first, then read the summaries. A team can also maintain a list of “sources we can cite,” so important decisions don’t rest on short social media posts alone.
This adjusts your entry points; it isn’t item-by-item review
The three questions take time to answer, and you can’t check every item carefully. This method is for adjusting entry points, not reviewing each post. A few judgments put an account into “avoid” or “keep,” and after that you don’t need to judge each thing it posts.
Excluding all AI-assisted content across the board also has a cost. Much valuable content uses AI for captions, transcripts, or translation, and uses like entertainment or gathering ideas, which don’t affect big decisions, don’t need the same strict standard. Strictness matters most for content that can affect decisions about health, money, the law, or work.
Next time you see an AI label, ask those three questions first. If none of the answers is clear, you can keep it out of your entry points even if the platform hasn’t given you an off button. If you use AI research tools to gather material, the sources they attach need the same check; see A Source Cited in an AI Research Report Shows Where It Read Something, Not That It’s True Scam texts are an extreme version of the same problem: however real the content looks, first ask what it wants you to do; see Scam Texts Look Real Now. Ask Whether You Can Skip the Link.
AI handoff card
Ask AI to help reshape your daily information diet This prompt asks where you most often meet low-quality content, uses three questions to sort your sources into avoid, skip, and keep, and suggests one adjustment for each platform today.
Every day I see a lot of low-quality content on social media, video platforms, and search, much of it AI-generated, and I want to reshape where my information comes from.
The method comes from this article: https://boosterminiclass.com/en/posts/ai-generated-content-filter-checklist/
Its view: an AI label only says content may have been made with AI; it won't filter for you. Ask three questions instead: does it have sources, is someone accountable for the judgment, does it state its limits? Then sort sources into avoid, skip, and keep.
Ask me 3 questions first:
1. On which platforms and accounts do I most often meet content that tires me or that I'm unsure I can trust?
2. What topics is it usually about, and could it affect my decisions about health, money, the law, or work?
3. Which sources do I currently trust and want to spend more time on?
Based on my answers, give me:
- A list sorting the sources or accounts I mention into avoid, skip, and keep, with which of the three questions each failed;
- One concrete adjustment for each platform that I can make today (for example unfollowing, tapping "Not interested," or building a list);
- Which topics I should check against official or original sources first, before reading summaries.
Don't judge whether specific accounts are genuine or accuse anyone. If information is missing, ask me first.
The list is only a starting point. Which sources are worth keeping is your decision.
References
- YouTube Official Blog: How we’re helping creators disclose altered or synthetic content — https://blog.youtube/news-and-events/disclosing-ai-generated-content/ (2024-03-18)
- Meta Newsroom: Our Approach to Labeling AI-Generated Content and Manipulated Media — https://about.fb.com/news/2024/04/metas-approach-to-labeling-ai-generated-content-and-manipulated-media/ (2025-10-23)
- Social Media Today: TikTok Adds Option To Limit AI Content in Feed — https://www.socialmediatoday.com/news/tiktok-adds-option-to-limit-ai-content-in-your-for-you-feed/805952/ (2025-11-19)



