Meta wants your website in the bargain
Meta's revised commercial terms extend crawling permissions to websites associated with a business. The question is what businesses receive in return, and how freely they can refuse.
Meta is asking businesses to put more into the bargain. Keep using its platforms commercially, and its revised terms seek permission to collect content from websites associated with your products or services. That includes material you have never posted to Facebook or Instagram.
Meta’s updated Commercial Terms apply in Australia and are scheduled to take effect on 30 October 2026 for commercial use of covered Meta products.
My view is that this is another significant overreach. A platform relationship should not become an expanding claim on the rest of your business.
Over the weekend, I raised this issue with the Attorney-General because I believed it needed immediate attention.
What the terms actually say
I compared Meta’s current commercial terms with the preview taking effect on 30 October 2026. Section 1 authorises automated collection from associated websites, periodic updates and the creation of “test credentials”.
The first version expressly included training Meta products among the permitted uses. After criticism, Meta narrowed the wording. The English (US) preview now ties crawling to advertising and business products, plus product integrity and security. The explicit training sentence is gone. That is a meaningful change; the crawling permission remains.
Meta describes the update as consolidation. This comparison uses the narrowed English (US) version and earlier generative-ad and catalogue provisions.
The wording you see may differ with the page’s selected language. On 19 October, English (UK) still included the earlier training sentence; English (US) showed the narrower wording above. You can compare them using the language selector at the foot of Meta’s preview page. I could access both from the same location, so I have not established a geographic split in permissions.
| Area | Before | Updated preview | What this means for marketers |
|---|---|---|---|
| Who grants crawling permission | Product-specific provisions, including generative ads and catalogues. | General commercial terms, also covering business Page management. | Review permissions across organic activity and advertising. |
| Which websites it covers | Ad and catalogue provisions centred on supplied URLs. | Associated websites and URLs made available through use of Meta services. | Consider research, guides and other content beyond campaign landing pages. |
| How content is used | Ad terms already covered AI improvement; catalogue terms supported ads and optimisation. | Advertising and business products, plus integrity and security; explicit training sentence removed. | Check each permitted use. The revision does not establish a blanket ban on AI use. |
| How businesses accept | Continued commercial use after notice meant acceptance. | Same mechanism retained. | Keeping campaigns or Pages running can carry contractual consequences. Assign someone to review changes. |
| Notice of future changes | No fixed minimum notice period in the commercial terms. | At least 30 days for material changes, with safety, security and legal exceptions. | Use the notice period to coordinate marketing, content, legal and web teams. |
The broader reach across commercial users is the issue. AI-related permissions and acceptance through continued use both existed before.
Meta says this is not about scraping news sites, crawling practices are unchanged and robots.txt preferences still apply. The narrower wording deserves acknowledgement. So does the need for clear, specific permission.
Agree, or leave your audience behind
Section 7 makes continued commercial use acceptance of the updated terms. Leaving is the stated alternative.
For a business that depends on Facebook or Instagram for enquiries, sales or customer contact, leaving can carry a serious commercial cost. Years of audience building cannot simply be transferred to another platform. The ability to close an account says very little about the freedom to negotiate.
The larger party sets the terms, changes them, and points to the exit as evidence of choice. The smaller party calculates what walking through it would cost.
That does not automatically make a clause unlawful. Australia’s unfair contract terms protections have specific tests, and courts decide whether a term fails them. My objection is commercial: dependence is being used to obtain a broader permission without a separately negotiated return.
The problem was always bigger than news
Australia has history here. Facebook temporarily blocked news in 2021 during the dispute over the News Media Bargaining Code. The Code sought to address bargaining imbalances between platforms and news businesses over payment for news made available on their services.
I was not a supporter. I thought payments from platforms would not, on their own, resolve publishers’ underlying business-model problems. The loss of classifieds and the shift in advertising online had already undermined the economics of news. Payments could help fund journalism; I doubted they amounted to a sustainable strategy for the businesses producing it.
I also believed Facebook benefited from the audience signals generated around news. Clicks, shares and reactions could help it understand people’s interests and refine advertising audiences. The ACCC’s 2019 inquiry examined the advertising value of user data and the richer datasets available to platforms. News was one source of those signals.
The connection to today’s terms is who gets to set that value exchange. The earlier dispute concerned news made available through platforms; these terms seek permission to collect content from businesses’ own websites. My concern is how much value the platform can capture, and how much say the business supplying the content has in the bargain.
Access becomes precious on the other side
The timing is striking. Search suppliers report Google making automated collection harder. BrightLocal describes /goto redirects obscuring destination URLs. Nozzle reports normal-looking result pages with anomalous substitutions, which it calls poisoning. Those are supplier observations and interpretations; they do not establish Google’s intent to serve decoys.
Google’s case against SerpApi alleges circumvention of protections around licensed third-party content in search results. It is not a settled finding that Google owns the web: the initial complaint was dismissed in July, and Google filed an amended complaint in August.
Large platforms understand the value of controlling access to their services. Website owners deserve a meaningful say over theirs.
The value exchange keeps shrinking
In The agentic layer, I argued for permission, attribution and a return of value to the organisations supplying the web’s useful material. That thinking underpins my work on Odexa.
The original wording showed why I worry about an AI content land grab through platform agreements. Meta’s revision narrows that concern; it does not settle the broader question of how freely businesses can negotiate over content rights.
Fair access needs a clear purpose, meaningful choice and a return the content owner can assess. Broad permissions followed by clarification after a backlash do little for Meta’s credibility.
An acceptance recorded in a terms database is a poor substitute for a fair bargain.
Tell your Meta representative
If you have a Meta representative, contact them and express your concern about the new terms. Ask which website content Meta intends to collect, how it will use it, and whether you can refuse those uses without losing access to the services your business relies on. Ask for a clear answer in writing.