How to remove a news article about you, layer by layer

01Where the article actually lives

Most people start with the search result, because that is where they keep seeing the article. The search result is only a pointer. Behind it sit up to five separate copies, and someone different holds the delete button for each of them.

The order matters more than the wording. A search engine asked to drop a result while the original page is unchanged will usually refuse, and if it does act, the result can return once the page is crawled again. Google states that its outdated content tool only updates the search result and does not remove a page from the web. So the work runs from the source outwards, never the other way round.

02What kind of article it is

The second question is what is actually wrong with the article, because that decides which door is open. People tend to assume that only false articles can be touched and that accurate ones are permanent. Both assumptions are half right.

In the UK the Editors' Code of Practice, which IPSO enforces, says a significant inaccuracy, misleading statement or distortion must be corrected promptly and with due prominence, and that a fair opportunity to reply to significant inaccuracies should be given when reasonably called for. Those two duties are the strongest lever a private person has over a UK newspaper. Neither of them obliges the newspaper to delete anything.

03Write to the publisher first

Almost every route, the UK regulators included, expects you to have gone to the publication first. A request that reaches the right desk and asks for something specific gets an answer far more often than a demand to take everything down.

Send it to the corrections or readers' editor address rather than to the reporter, so that the request is logged. Identify the article by headline, date and URL, and quote the exact sentences at issue. Then ask for the narrowest change that solves the problem: removing your surname, removing a photo, rewriting a headline, adding an update, or adding a noindex tag so the page drops out of search engines while staying on the site.

Some US newsrooms run formal review programmes. The Boston Globe launched its Fresh Start initiative so that people can appeal their presence in older stories. The Globe says it may update a story or remove it from Google search results, reviews each request case by case, and is not in the business of rewriting the past. That is the realistic tone of most publishers' answers: changes are possible, deletion is the exception. If nothing comes back, send one reminder that quotes your first request and its date, then move to the regulator or to the next layer rather than writing again and again.

04UK: complaints to IPSO and Impress

If a UK publication refuses or does not reply, the next step depends on who regulates it. Newspapers and magazines that have signed up to regulation by IPSO go to IPSO; publishers regulated under the Impress scheme go to Impress. Check the regulator's member list before filing, because neither can deal with a title outside its scheme.

IPSO says you will normally need to complain within four months of publication, and that for online articles it may be able to investigate up to twelve months after publication, depending on the circumstances. It normally passes the complaint to the publication first, with a default period of 28 days for the two sides to resolve it, and says it can also help to mediate between you and the newspaper; that stage can be skipped if you have already corresponded with the publication without result. If the complaint is upheld, the Complaints Committee can decide that a correction or an adjudication setting out its findings should be published.

Impress requires you to contact the publication first and expects the publisher to resolve a complaint within 21 calendar days. It accepts complaints made within four months of becoming aware of the content, and no later than twelve months after the original publication.

05US: what the law will not do

In the United States the editor's discretion is close to absolute. The Supreme Court struck down a state law that forced newspapers to print a reply, holding that it intruded into the function of editors, so there is no legal right of reply to rely on. A US publisher that declines to change an accurate story has, in most cases, the last word on its own page.

Search does not fill the gap. Google's personal information policy covers narrow categories such as home addresses, phone numbers, ID numbers, bank details, images of signatures or ID documents, private records and login credentials. Google adds that if content is about something important or newsworthy, like the news, it might not take it down. An accurate article about you fits none of those categories.

What remains is the publisher's own policy, a defamation claim where the article is false and damaging, and accurate material about you that ranks for your name. A claim belongs with a lawyer licensed in the state concerned. The practical job before that is a clean record: every version, every date, every copy.

06Acquitted, dropped or closed

The most common request from private individuals is an arrest or charge story that was never followed by a report of the outcome. The fix is not a deletion demand. It is new information the editor did not have when the story ran.

• Send the court record or the prosecutor's decision showing acquittal, dismissal or no further action.
• Ask for an update at the top of the article, or a follow-up report, and for the headline to reflect the outcome.
• If you were a private person, ask for your name and photo to come out of the archived version.
• If the publisher will not change the text, ask it to stop the page being indexed by search engines.

Under European data protection law the outcome also opens the search route. Google's published criteria say that information which became outdated after publication, giving the example of a case later overturned on appeal, may be delisted, and that a court ruling in your favour receives significant deference. Send the court document with the request rather than describing it.

07Aggregators and syndicated copies

Aggregators do not edit the stories they carry. Google News tells anyone who did not publish the content to contact the publisher directly, because the publisher can remove it from its site and ask its own team to contact Google. Google says a publisher fixes an error by republishing the page without it, that removal can take up to a few days, and that incorrect content still showing on Google News can be reported through a contact form with the exact headline.

News apps and portals that take publisher feeds work the same way in practice: the change has to happen at the source. Once the publisher acts, check every syndicated copy separately, because a copy stored under another site's address does not always follow the original.

08Copies on blogs, forums and scrapers

Once the original changes, the copies become the problem. Whole articles pasted into blogs and forums often climb in search results after the source page disappears, because they are now the only version left.

The obvious shortcut does not work. A US copyright takedown notice must carry the signature of a person authorised to act for the owner of the copyright, and the person named in an article is neither its author nor its publisher. The publisher can send one, and some will once they have corrected or removed their own page.

The working ground is the content itself: a privacy intrusion, a false statement, or a breach of the site's own rules. Each report should name the rule, quote the passage and attach the corrected or removed original. A copy of a page the publisher has already corrected is far easier for an operator to take down than a copy of a page that still stands.

09Search results: delisting and its limits

When the source page is gone or has changed, Google's Refresh Outdated Content tool asks it to update the result. Google says the tool will not remove a page that still exists and has not changed. If the page was changed rather than removed, Google drops the old snippet and refreshes the result the next time its crawler visits.

Where the article is accurate and still online, delisting is the only search route, and it exists under European data protection law. Google weighs whether a result is inaccurate, inadequate, irrelevant or excessive against the public interest. It is less likely to delist information about a public role and more likely to delist purely private matters, and for journalistic material it considers the outlet's editorial judgement, particularly for recent articles. A delisting covers searches on your name only.

The right to erasure continues in UK law. If Google declines, you can ask a data protection authority to review the decision; the ICO says you should complain to the organisation first and then to the ICO if you remain dissatisfied. Google also says it aims to notify the website owner where possible when a page is delisted, so assume the publisher will learn that a request was made.

10Archives, and the order of work

Archive snapshots and reposted screenshots survive every other step. The Internet Archive takes exclusion requests at info@archive.org and asks for the URLs, the time period to be excluded and, where relevant, the period during which you controlled the site. It reviews each request and does not promise a result in advance. Screenshots on other sites are handled like any other copy.

1. Fix the dates: publication, when you found the article, and any court outcome.
2. Map every layer: original, aggregator copies, reposts, search titles, images, archives.
3. Go to the publisher with the narrowest request that solves the problem.
4. If refused in the UK, complain to IPSO or Impress inside their time limits.
5. After the source changes, report each copy with the change attached.
6. Refresh or delist search results, then search your name again, images included.

SARAZIM LLC runs this sequence end to end: the layer map, the requests to publishers, operators and search engines, and monitoring for reposts afterwards. Face-recognition detection finds where a photo from the article has spread, starting from two or three pictures of you, and the company keeps no copies of the material, only metadata and hash values. It averages more than 52,142 removals a month across 914+ platforms. It is not a law firm; where a legal claim is needed, it hands your lawyer a complete record.