A client sends you a link. A news story, a mugshot site, a people-search page with their home address on it, a Reddit thread calling the business a scam. The question is always the same: can you make this go away?

The answer depends entirely on what kind of page it is, and the most expensive mistake in this work is skipping that diagnosis. Teams routinely spend a year burying content that was legally removable, or run a takedown campaign against a public record that was never coming down.

Negative content removal has also become more urgent than it was two years ago, because the failure mode changed. When a result sat at position eight it was effectively invisible. When an AI assistant retrieves that same page to build an answer, ranking position stops protecting anyone.

Three outcomes get called “removal” and only one of them is

Before touching a single URL, get precise about what you are actually promising.

• Removal. The source page no longer exists. The publisher deleted it, the platform took it down, or a court ordered it gone. This is the only outcome that also stops the page from feeding AI answers.
• Deindexing. The page is still live, but Google no longer returns it. Anyone with the link can read it, and crawlers, scrapers and retrieval systems can still reach it.
• Suppression. Nothing was removed or deindexed. You have published enough stronger material to push the page below where people look.

Suppression is still legitimate work and often the only realistic play. It is simply the weakest of the three against AI surfaces, because a model assembling an answer does not care which position a source held. Pew Research Center’s browsing analysis found that when an AI summary appears, people click a traditional result only about 8% of the time, roughly half the rate of searches without one.

Adoption raises the stakes. BrightLocal’s 2026 Local Consumer Review Survey found 45% of consumers now use tools like ChatGPT for local business recommendations, up from 6% a year earlier, which is why checking what AI says about your locations belongs on the same checklist as reputation work.

Triage first: three questions per URL

Run every problem page through these before committing budget.

1. Is the page still live, or has it already come down and simply lingers in caches and copies?
2. Can you point to a specific factual error, rather than a fact that is merely unflattering?
3. Is there a court order, dismissal or expungement attached to it?

A yes on the second or third question usually means a removal path exists. A no across all three, particularly on a legitimate news or government site, means suppression is your realistic strategy. Write the verdict next to each URL so nobody re-argues the same case in three months.

News articles: the publisher is the lever, not Google

Google will not deindex a legitimate news article because someone dislikes it. Asking is not a strategy. The realistic paths, ordered by how often they work:

• Factual corrections. Most reputable outlets run a formal corrections process. A correction does not remove the article, but it often reframes the claim doing the damage. Use the outlet’s corrections or editorial standards page rather than a generic contact form, and lay out the error with dates, names and documents.
• Unpublishing requests. More viable than most people assume at local and regional outlets. NPR has documented how programs like The Boston Globe’s “Fresh Start” and Cleveland.com’s “Right to Be Forgotten” let people petition to have old stories updated, anonymized or delisted, typically where the subject is not a public figure. Policies vary, so check before assuming no.
• Defamation action. Viable where content is false and provably damaging, slow and expensive where it is merely unflattering. A court order is the one instrument that reliably delivers both removal and deindexing, and that is a decision for a qualified attorney.
• Syndication cleanup. One article often lives across a dozen syndication partners, aggregators and scrapers. Killing the original and leaving the copies achieves almost nothing, and the copies are frequently what AI systems cite.

Budget for one procedural detail: even after a page comes down, Google can serve the cached listing and snippet for weeks. Google’s Outdated Content Tool is built for that gap, so submit the refresh request yourself rather than waiting on the crawler.

Mugshots and court records: more removable than people assume

The law has moved here, and it has moved in favor of the subject.

• Mugshots. The clearest case. As legal publisher Nolo outlines, many states now make it illegal for mugshot sites to charge removal fees, and both Google and the major payment processors have acted against pay-to-remove operators. Where charges were dropped, dismissed or expunged, requests carry real leverage and the sites know it. Search your state’s mugshot removal law before contacting anyone.
• Sealed or expunged records. These should not be published at all. If a record has been expunged and a site still displays it, you have a claim rather than a request. Verify the expungement is complete first, because the paperwork often is not.
• Records that remain legitimately public. The hard case. Removal is unlikely and suppression is the working strategy, though aggregator sites reproducing the record usually maintain their own removal processes separate from the court.

Sequence matters more than effort here. Expunge or seal first where possible, then pursue removal; doing it in reverse means re-litigating every request with weaker documentation. Pull the dismissal paperwork or expungement certificate before submitting anything, because a request backed by a document is far harder to ignore.

Data brokers: procedural, endless, and largely solvable

Home addresses, phone numbers, relatives and property records surface through data brokers, the most systematically solvable category on this list. The California Privacy Protection Agency’s public registry alone lists more than 500 brokers, and under the state’s Delete Act, California residents can use the DROP platform to send one deletion request to every registered broker, with brokers required to begin processing those requests in August 2026.

Google’s Results About You is the other free tool worth running. It monitors your name, address and phone number in search results, and a 2026 update extended it to exposed government ID numbers such as driver’s license and Social Security numbers.

Two things make brokers different: the volume runs into hundreds of sites rather than a handful, and records reappear whenever a broker refreshes from public sources. That is why one-time opt-outs fail. Treat it as recurring maintenance, not a project with an end date.

Reddit and forum threads: rarely removable, increasingly loud

Reddit will not remove a thread for being unflattering. Moderators sometimes will where a post breaks subreddit rules, and posts containing personal information or clear harassment can be actioned under platform policy. Start with the in-app report, then message moderators through modmail citing the exact rule broken. Specific policy citations get answered; general complaints do not.

What changed is not removability but reach. A thread with thirty upvotes that never cracked page one now gets scraped into aggregators, quoted in roundups, and pulled into retrieval when someone asks an assistant about your brand. Google’s “Discussions and forums” module elevates the same threads that language models disproportionately cite, which is one route by which AI answers end up recommending your competitors. Before spending more time on an original post, check whether it has already been quoted or screenshotted elsewhere; if the citations point at downstream copies, the thread itself is not your problem.

Verify against AI answers, not rank checks

The old verification was a rank check, and it no longer tells you what you need to know. A page can be removed and deindexed while the claim inside it keeps appearing in AI answers, carried by copies you never found or absorbed into a model before the takedown landed. Pew Research found roughly one in five Google searches already triggers an AI summary, rising to around 60% for question-style queries, which is exactly how people search a business name plus “reviews” or “scam.”

A workable post-removal check:

1. Run the queries that surfaced the problem against ChatGPT, Perplexity and Google AI Overviews, not just Google Search.
2. Record every cited URL. Cited sources are the working set shaping the narrative.
3. Compare against your removals. Anything left is unfinished work with an address.
4. Repeat weekly. Outputs vary between runs, so one clean check is a sample, not a result.

That citation list converts a vague sense of “people are saying bad things” into a finite, ranked list of pages. It doubles as a visibility audit, which is why it pairs with our AEO playbook for earning AI citations and with what we found about how ChatGPT decides who it recommends before it searches.

Frequently asked questions

How long does negative content removal take?

It depends on the path. Broker opt-outs often process in days to weeks, though re-listing makes the work recurring. Publisher unpublishing decisions take weeks. Legal routes, including expungement followed by removal requests, are measured in months. Suppression usually needs several months of sustained publishing before results hold.

Should I handle this in-house or hire a specialist?

Broker opt-outs and platform policy reports are reasonable in-house tasks if someone owns the recurring maintenance. Expungements, defamation and widely syndicated content are harder to run alone, and a sequencing mistake can cost you leverage permanently. If you hire, ask which of the three outcomes is being promised and get it in writing.

Will removed content come back?

Sometimes. Broker listings reappear when datasets refresh, and removed articles persist through syndicated copies. Ongoing monitoring matters as much as the initial takedown.

Does deindexing stop AI systems from using a page?

Not reliably. Deindexing changes what Google returns to human searchers. The URL stays live and reachable, so retrieval systems, scrapers and aggregators can still pick it up. Only genuine removal at the source closes that door.

Diagnose before you spend

The first question in every case is diagnostic, not tactical. Is this page removable, deindexable, or only suppressible? Most wasted budget comes from running suppression against content that had a removal path, or attempting removal on content that never had one.

Three next steps: sort every negative URL into one of the three outcomes, check for syndicated copies of anything you plan to remove, and run your branded queries through the AI surfaces to build a citation list. Get the category right first. The tactics are the easy part.