How to request a Google Search removal in Australia

A Google result is not the page. It is a card Google chose to print on top of the page. Australians who write to “Google support” as if it were an editor will wait forever. The company runs a set of narrow products that hide, trim or de-index a URL. None of them unpublished the Age, a WordPress blog, or a Facebook screenshot that has already been scraped.

If you remember one distinction from this guide, make it this one: source takedown versus search takedown. Do the first wherever you can. Do the second so that the corpse stops appearing on the first page of a name search. The rest of the platform takedowns hub covers the source. The search and archives hub covers caches, sitelinks and the copies Google does not control.

What Google will actually consider

Google’s public removal tools for people in Australia sit in a handful of buckets. They change labels, but the buckets have been stable for years.

Personal information that is already out of date or exposed. Phone numbers, home addresses, email addresses, identity-document images, financial account numbers, handwritten signatures, and certain intimate images can be nominated through Google’s personal-information forms. The page usually stays up. The snippet and the index entry are what you are asking to change.

Doxxing and explicit “people-search” packaging. Where a site is in the business of assembling a dossier — name, suburb, relatives, old phone numbers — Google has become more willing to de-index the packaging even when each field is separately public. That is not a general privacy right. It is a product decision, and it is why data-broker opt-outs still matter: Google will not stay ahead of a broker that republishes under a new URL every month.

Legal removals. Court orders, including injunctions and some defamation judgments, can be submitted through Google’s legal investigations form. You need the order, the URL, and a clear statement of what the order requires. A solicitor’s letter that is not an order is lobbying, not a legal removal. The defamation guide explains how you get to an order in the first place.

Copyright. If you own the copyright in the work being shown — your photograph, your article, your video — Google has a copyright complaint flow. If you merely dislike an article about you, copyright is the wrong door. We wrote a whole piece on that mistake: copyright takedowns versus reputation takedowns.

Child sexual abuse material and non-consensual intimate images. Do not start with a journalist and do not start with this desk. Use the platform, use eSafety, and then use Google’s sensitive-image products. See non-consensual intimate images and when a child is in the footage.

What Google will not do, even if you ask politely and often: weigh in on a one-star review that is merely rude; unpublish a news article because you have moved on; apply the European delisting test to an ordinary Australian name search; or promise that Bing, DuckDuckGo, or an AI Overview will follow.

Prepare the file before you open the form

Most failed submissions are incomplete, not unconvincing. Build a short dossier.

1. The exact URL, copied from the result, not from a mobile share-sheet that has stripped the query string. If Google is showing a cached copy, capture both the live URL and the cache URL.
2. A screenshot of the result as it appears in an incognito window in Australia, with the date visible. Search is personalised. You need the public card, not the one your logged-in account has already demoted.
3. A screenshot or PDF of the underlying page, in case the publisher moves it while you wait.
4. The legal or policy hook in one sentence. “This URL displays my current home address.” “This URL is named in paragraph 4 of the attached Federal Court order.” “This URL shows an intimate image of me published without consent.” If you cannot write that sentence, you do not yet have a Google job. You have a source job.
5. Proof you are the person. A redacted licence or a utility bill is usually enough for personal-information products. For a company name, you will need to show you are authorised to speak.

Keep the dossier off the cloud account that is already compromised. The first 48 hours after a leak is the wrong moment to upload a passport to a random “reputation” vendor.

The forms, in the order we use them

Start at Google’s Help Centre removal tools, not at a Twitter complaint. The interface is deliberately boring.

Outdated content. If the page itself is already gone or heavily updated and Google is still quoting the old snippet, the outdated-content tool is the lightest touch. It is not a reputation product. It is a freshness product. Use it when the publisher has done the right thing and the index has not noticed.

Personal information. Use this for contact details, identity documents, and the doxxing-style compilations. Name every URL. Google does not “clean a name”; it decides URL by URL. If the same address sits on twelve broker pages, you will file twelve times, then file again when the broker clones the path. That is also why you should run the broker opt-out programme in the same fortnight.

Legal. Attach the order. Quote the operative paragraph. List every URL the order covers, including obvious mirrors if they are named. If the order is suppressed or anonymised, say so in the cover note so the reviewer does not paste your name into an internal ticket that later leaks.

Images. There is a separate flow for images that appear in Google Images, including intimate images and photos of you attached to an unrelated scandal. An image takedown in Search does not remove the file from the host, and it does not stop a new scrape from creating a new URL.

After you submit, you will receive an automated receipt. That is not a decision. Decisions arrive later, often as a short email that names the URL and says removed, denied, or more information required. Keep the thread. If you later need a solicitor to write a concerns notice to the publisher, the Google correspondence is useful context, not a substitute.

What “removed” means on the results page

A successful personal-information or legal removal typically means the URL no longer appears for the queries Google associates with the complaint. It does not mean:

• the page is gone;
• the cached copy is gone on the same day — see caches, sitelinks and Knowledge Panels;
• Bing, Yahoo, DuckDuckGo, Yandex, or a Chinese index have followed — see the other indexes;
• Gemini, ChatGPT, or Google’s own AI Overview have stopped summarising the old text — see what AI Overviews do with harmful results;
• the Wayback Machine has dropped its copy — see archive removals.

It also does not mean the Knowledge Panel on the right of the results has been rebuilt. Knowledge Panels are assembled from Wikipedia, Wikidata, news graphs and various structured sources. A single de-indexed URL will not empty a panel that is drawing from a Wikipedia paragraph you have not touched.

The Australian legal overlay

Google will honour a clear Australian court order that tells it to stop making a URL available in Search. Getting that order is a different profession. Defamation remains the main civil route for a false allegation that is causing serious harm. The Online Safety Act is the main administrative route for adult cyber-abuse and image-based abuse; eSafety can give removal notices to services, which is often faster than suing Google itself. Read what the eSafety Commissioner can actually order before you assume a Federal Court writ is the first move.

Privacy complaints under the Privacy Act are usually aimed at an APP entity that holds your information, not at Google-as-search-engine. There are edge cases, and they are for a solicitor who has read the current OAIC guidance, not for a contact form on a journal.

One more Australian peculiarity: the result you see in Sydney is not always the result a London recruiter sees. If the damaging card is appearing for overseas searchers, say so in the legal cover note and attach both screenshots. Google’s indexes are not a single drawer.

When the form will fail, and what to do instead

The form fails when the content is a news article that is still live, a review that is merely negative, a political claim, or a forum post that is awful but not personal information. In those cases the work moves:

• to the publisher’s corrections desk, or to a concerns notice, as in news archives;
• to the review product’s own policy, as in Google Reviews;
• to the platform that hosts the thread, as in Reddit and forums;
• to a long programme of source cleanup, after which the outdated-content tool becomes useful.

Agencies that promise to “wipe Google in 48 hours” are selling a product Google does not offer. Some of them file the same personal-information form you can file, then invoice you for the receipt. There is no shame in paying someone to do tedious paperwork. There is shame in believing the paperwork is a spell.

A working sequence we recommend

1. Capture the public results page and the underlying URLs.
2. Take the source down, or start that process, using the relevant platform guide.
3. File Google personal-information or legal removals for every URL that qualifies.
4. File the same URLs with Bing and any other index that is actually referring people.
5. Wait, then use the outdated-content tool on anything the publisher has already fixed.
6. Recheck AI Overviews and the Knowledge Panel a fortnight later, not the same afternoon.
7. Diary a reminder in 90 days. Brokers and scrapers come back.

If you want the desk to look at a general question about the process, use contact. Do not send the intimate image, the court file, or a zip of the leak. We publish method. We do not run your case.