Great Disappearance Acts Generative Search and Shadow Banning
This paper argues that generative search and shadow banning threaten the open web's sustainability and democratic character by diverting traffic and suppressing visibility, necessitating enhanced regulatory interventions and algorithmic transparency to ensure fairness, accountability, and diversity in online expression.
Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
The Big Picture: The Internet is Getting "Sticky" and "Silent"
Imagine the internet used to be a giant, bustling town square. In this square, anyone could set up a booth (a website), share their news or art, and people would walk by, stop to chat, and maybe buy a snack (click an ad or subscribe). This was a fair, open place where small vendors could thrive alongside big stores.
The author argues that two new, invisible forces are quietly dismantling this town square. They aren't tearing down the booths; they are just making it so no one can find them, or so no one bothers to walk inside anymore.
Act 1: The "All-Knowing Librarian" (Generative Search)
The Old Way:
Imagine you go to a library and ask a librarian, "Who wrote The Great Gatsby?" The librarian points to a shelf and says, "It's in Book A, Section 3." You have to walk over, pull the book, and read it yourself. The library gets traffic, and the book gets attention.
The New Way (Generative Search):
Now, imagine that librarian gets a super-brain upgrade (powered by AI). You ask the same question, and instead of pointing to a shelf, the librarian instantly recites the answer: "F. Scott Fitzgerald wrote it in 1925."
The Problem:
Because the librarian gives you the answer right there at the desk, you never walk over to the bookshelf.
- The Result: The books (websites) get no visitors.
- The Consequence: If no one visits the bookshelf, the library can't sell tickets or ads to the people walking by. The small book publishers (independent creators and small businesses) go out of business because their "traffic" (visitors) has vanished.
The Legal Mess:
The librarian is reading all the books to learn how to answer your questions. The authors of those books are furious. They say, "You are reading our work, summarizing it, and giving the answer away for free, stealing our customers."
- Copyright: Did the librarian steal the story?
- Unfair Competition: Is it fair for the librarian to use the authors' hard work to build a business that doesn't pay them?
- The "Semi-Plagiarism": Sometimes the librarian says, "Here is the answer," and lists the book titles at the bottom in tiny, hard-to-see text. It's not a direct theft, but it's close enough to be suspicious.
Possible Fixes:
The paper suggests the librarian could pay the authors a fee (licensing) or use a new type of "AI Agent" that acts like a middleman to ensure the authors get paid when their info is used. But right now, the big tech companies are mostly just taking the content without paying.
Act 2: The "Silent Bouncer" (Shadow Banning)
The Scenario:
Now, imagine you are at a party (a social media platform like YouTube or TikTok). You post a video. You aren't breaking any rules, and you aren't stealing anyone's music. But suddenly, your video stops getting views. No one sees it, but you aren't told why. You aren't kicked out; you are just ignored.
The Mechanism:
This is called Shadow Banning.
- The Bouncer's Logic: The party bouncer (the algorithm) is scared. If a copyright owner (like a big music label) complains, the bouncer doesn't want to get sued.
- The "Safe" Move: Instead of deleting your video (which might be a mistake if your video was actually fair to use), the bouncer just turns down the volume on your microphone. They hide your post in the deep corners of the party so no one finds it.
- The "Twilight Zone": The bouncer doesn't know if your video is actually illegal or just legal but unauthorized. So, to be safe, they suppress everything that looks even slightly suspicious.
The Result:
- Chilling Effect: People stop sharing things because they are afraid their posts will be silently buried.
- No Appeal: If you ask, "Why is no one watching my video?" the bouncer says, "The algorithm decided." There is no clear way to argue your case.
- Dark Data: This suppression happens in the shadows. You don't know it's happening until it's too late.
The Regulations:
The paper looks at how China and the EU are trying to fix this.
- China: Has rules saying algorithms must be transparent and fair, but it's hard to enforce when the "black box" of the algorithm is so complex.
- EU: Has new laws (the AI Act) trying to categorize these risks, but the paper argues they still don't go far enough to protect the right to speak freely.
The Conclusion: A Centralized, Controlled Square
The paper concludes that the internet is shifting from a decentralized town square (where many small voices can be heard) to a centralized mall controlled by a few giant landlords (Google, Meta, etc.).
- Generative Search steals the traffic that keeps small businesses alive.
- Shadow Banning silences voices to protect the landlords from legal trouble.
The author warns that if we don't fix this with new laws and transparency, the internet will lose its soul. It will stop being a place for free expression and innovation and become a tightly controlled space where only the biggest, richest players survive, and the rest of us are just passive consumers of answers we didn't have to work for.
In short: The internet is becoming a place where the "answer" is given to you for free (killing the source), and if you try to speak up, you might be quietly muted without a reason. The paper calls for rules to make the system fair, transparent, and open again.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.