Pinterest notified advertisers on September 4, 2026 that a rewritten set of Advertising Guidelines takes effect on November 12, adding a standalone prohibition on scams and deceptive commerce that names synthetic media, scam tooling and fictitious shopfronts as advertising categories for the first time.

In Short

Pinterest is rewriting the rules that decide which ads can run on its platform, and the new version takes effect on November 12, 2026. If you buy ads on Pinterest, the change adds a section on scams that did not exist before, covering AI-made product claims, fake giveaways and shopfronts that do not exist. It also rewrites the contracts sitting behind Pinterest's television and off-platform ad products, so the paperwork attached to an account opened last year is not the paperwork that will govern it in November.

A notice sent 69 days ahead of the deadline

The notification arrived by email at 05:05 on September 4, 2026, sent from Pinterest's recommendations address. Sixty-nine days separate that date from the effective date. The message covers a bundle rather than a single document: Terms of Service, Privacy Policy, Community Guidelines, Advertising Guidelines, Merchant Guidelines, the Advertising Services Agreement and the Ad Data Terms incorporated within it.

According to Pinterest, the Terms of Service rewrite was intended to make the documents easier to understand and to set out the process that applies when an advertiser or user has a dispute with the company. The consent mechanism is passive. Continued use of Pinterest and its services after November 12 constitutes agreement to the updated terms, according to the notice.

Two details in the notice do not match the underlying documents. The subject line describes the change as an update to the Business Terms of Service, while the body refers only to the Terms of Service without the business qualifier. More substantively, the email names Community, Advertising and Merchant Guidelines as the three guideline sets revised in connection with scams and deceptive goods. A fourth document, the GenAI Acceptable Use Guidelines, carries the same November 12 effective date and contains a material addition, but does not appear in the email at all.

Pinterest has published both versions of the affected policy pages in parallel. The live Advertising Guidelines page carries a banner pointing readers to a preview of the replacement, and the preview page carries the reciprocal pointer back. The same arrangement applies to the GenAI document. That makes a line-by-line comparison possible before the switchover rather than after it.

Thirteen prohibitions that did not exist as a category

The clearest structural change in the Advertising Guidelines is a new heading, "Scams, deceptive goods and practices", positioned between unacceptable business practices and weapons and explosives. Its opening paragraph bars advertising for anything that facilitates or promotes, in Pinterest's phrasing, "dishonest commerce, exploitative schemes or deceptive practices", and names exaggerated claims, intentional misrepresentation, misleading pricing, bait-and-switch tactics and deceptive redirects as examples.

Thirteen bullet points follow. Eight of them are relocations: fake online tech support, hidden costs and bait-and-switch pricing, homework and essay writing services, sale of social media followers, negative option billing, unrealistic or exaggerated claims, get-rich-quick and multilevel marketing schemes, and ads that intentionally show a different product than the one being sold. All eight sat in the unacceptable business practices list in the current text.

Five are genuinely new. Pinterest will prohibit products, services or offers misrepresented through synthetic media or manipulated content. It will prohibit tools or templates intended to facilitate, automate or scale scams, which reaches the supply side of fraud rather than the fraudulent offer itself. It will prohibit fake or deceptive offers, grants, prizes or giveaways used to mislead people. It will prohibit fictitious or non-existent products, services or shopfronts. And it will prohibit impersonation or social engineering used to deceive people or misrepresent an identity, brand or affiliation.

Two of the relocated items were also widened in transit. The current ban on companies offering homework or essay writing services becomes a ban on services that enable academic dishonesty, with tutoring still permitted. The current ban on companies that sell social media followers becomes a ban on services that artificially inflate engagement metrics, with follower sales named as one instance rather than the whole prohibition.

The email describes these guideline revisions as clarifications of existing policy. The comparison shows five prohibited categories that have no counterpart in the current text and two that have been broadened from named business models to described behaviours.

The residual list, regrouped

With eight items moved out, unacceptable business practices shrinks from thirteen entries to six, arranged under three headings that the current version does not use. Predatory financial products holds binary options, payday loans and all cryptocurrency products and services except those the Financial Products and Services Policy permits. Exploitative schemes holds penny and bidding fee auctions. Legal system exploitation holds plaintiff recruitment services and bail bonds.

Binary options is the one addition. The instrument appears nowhere in the current Advertising Guidelines, and its arrival puts Pinterest alongside platforms that closed the category years earlier. Everything else in the six-item list carries over verbatim.

Promotional mechanics under the same heading are untouched. The 90 percent discount ceiling, the ban on sweepstakes and trade-in offers, the requirement that promo codes be generally available and reflected at checkout, and the four custom promotion conditions all survive the rewrite unchanged.

Targeting language loses its closed list

A smaller edit sits in the protected groups passage, and it changes the shape of the rule rather than its subject. The current text states that in certain markets, credit, employment and housing ads cannot use actalike, age, gender or postcode targeting options, or exclude protected groups. The replacement states that those ads cannot use actalike targeting, "certain demographic targeting (such as age, gender or postcode)", or exclude protected groups.

The parenthetical converts an exhaustive list into an illustrative one. Under the current wording, a demographic parameter outside the four named options falls outside the restriction. Under the replacement, the restricted set is defined by category rather than enumeration, and Pinterest determines what else belongs inside it. The three ad categories are unchanged, as is the note that campaigns mixing prohibited and permitted criteria may be approved without serving on the prohibited ones.

The rest of the targeting architecture survives intact. Eight sensitive categories remain closed to targeting, covering health conditions, race and ethnic origin, personal hardship, religious belief, sexual behaviour and gender identity, criminal allegations, trade union membership and political affiliation, alongside the prohibition on targeting anyone under 18 and on using proxies for any of the eight. The nineteen-item list of information advertisers cannot request through lead ads, running from precise geolocation to military status, is identical in both versions. So is the six-month deletion deadline on any Pinterest data tied to a unique identifier, and the ban on lead ads for prescription drugs, over-the-counter products, health apps and telehealth.

One rule gets looser

Landing page standards contain the only visible relaxation in the document. The current text bars destinations that "have pop-up ads or start automatic downloads on someone's device". The replacement bars destinations with non-dismissible pop-up ads and automatic downloads.

A dismissible interstitial on a landing page is a policy violation until November 12 and a permitted experience afterwards. The other four landing page conditions, covering gated content, forced redirects, made-for-ads destinations and inconsistent experiences, are unchanged, as are the eleven editorial quality rules governing image resolution, text overlay, mimicked Pinterest interface elements and low-quality user-generated content.

The GenAI document acquires a penalty

The GenAI Acceptable Use Guidelines gain one sentence, added to the end of the purpose and applicability section. Abuse of Pinterest's generative AI technology may result in account enforcement, including "loss of access to GenAI features, or account termination or suspension", according to the updated text.

That sentence has no equivalent in the current version. Until now the document described obligations across three categories of prohibited use, dangerous or illegal, misleading or deceptive, and explicit, hateful or violent, without stating what happens when they are broken. The three lists themselves are unchanged, down to the individual bullets on jailbreaking, reverse engineering, data scraping, impersonation, spam and phishing, non-consensual intimate imagery and minor safety.

One cross-reference was renamed. Where the current text points readers toward guidelines on "Prohibited and Regulated Products or Practices", the replacement points toward "Dangerous goods and activities". The remaining pointers, covering site security, intellectual property, impersonation, misinformation, spam, advertising, sexual content, hateful activities and violence, are unchanged.

The document remains silent on machine-generated advertising creative specifically. Its scope is the use of Pinterest's own generative tools, not the provenance of assets uploaded from elsewhere, and the new scam section in the Advertising Guidelines is where synthetic content is addressed on the paid side.

Television and offsite inventory enter the contract stack

The commercial substance of the notice sits in one paragraph about agreements rather than guidelines. The Advertising Services Agreement now incorporates Pinterest's Offsite Ad Terms by reference. The Ad Data Terms inside it were revised, according to Pinterest, to reflect changes to its advertising products, specifically the ability to buy ads on connected TV and to extend Pinterest campaigns across other publishers' sites and apps.

That language describes infrastructure Pinterest spent the previous nine months assembling. The company acquired tvScientific under an agreement announced on December 11, 2025, and closed the transaction on February 17, 2026 for $465.1 million. By the first quarter, Pinterest had begun feeding its audience data and Taste Graph signals into tvScientific's buying platform, according to chief executive Bill Ready. An early home furnishings partner reported a 190 percent increase in incremental audience reach and a 159 percent increase in incremental sales after layering Pinterest audience data onto CTV campaigns, a figure supplied by a single disclosed partner in early testing.

Until this update, those products ran on contract language written for a platform that sold impressions on its own surfaces. The November 12 documents formalise the offsite case, where an advertiser's Pinterest campaign reaches inventory Pinterest does not own and data moves between Pinterest and third-party sellers. Pinterest had already opened its inventory to outside demand paths through ads.txt entries for Index Exchange and Criteo in January 2025.

The privacy reference moves

The Privacy Policy revision is described in a single sentence. Data collected through the service may be used to improve and personalise the experience on Pinterest and the ads users see off the platform, according to the notice.

That clause interacts with an advertiser obligation the Advertising Guidelines carry forward without change: advertisers are told not to collect personal data that is not covered in the Pinterest Privacy Policy, and to email the company's privacy legal address when unsure. The obligation is stable. The reference point it depends on is not, because the Privacy Policy is one of the documents being replaced. An advertiser data practice sitting inside the permitted boundary on November 11 is measured against a different boundary on November 12.

The measurement and audience-targeting rules themselves are untouched. Advertisers must still disclose data collection and obtain legally required consent, still tell site and app visitors about third-party sharing for behavioural advertising and about opt-out routes through Pinterest personalisation settings and the AdChoices site, still withhold data relating to sensitive categories or collected from children, and still refrain from sending customer lists built on iOS app or website activity. The Pinterest tag, audiences and app install campaigns all remain inside that perimeter, as does the underlying logic of ad personalization that the tag feeds.

Country rules carry over

The country-specific section, which runs from Albania to Vietnam and sets separate prohibitions and restrictions for each market, appears in both versions without substantive alteration. Alcohol thresholds, infant formula age cutoffs, pharmacy and telehealth bans, lottery permissions and language requirements are unchanged market by market. The trade sanctions list, naming seven territories where Pinterest ad products are unavailable, is also identical.

Advertisers operating in a single market therefore face no local change from this update. The changed text is all in the horizontal layers: what may be sold, how targeting restrictions are scoped, what a landing page may do, and what contract governs data moving off Pinterest properties.

Why this matters for the marketing community

Scam advertising has become a legal exposure rather than a moderation question. A consumer group sued Meta in April 2026 over deceptive advertising under District of Columbia consumer protection law, a complaint built on the argument that a platform which both sells advertising and writes the standards those ads must meet can be held to those standards. Reuters reporting cited in that case put roughly ten percent of Meta's 2024 advertising revenue against ads categorised internally as promoting scams. Google's 2025 Ads Safety Report described bad actors using generative systems to produce deceptive ads at volume, with detection systems built for earlier patterns struggling to keep up.

Pinterest's new prohibitions on scam tooling and fictitious shopfronts track that shift in method. Documented cases have shown AI-generated advertising carrying misleading product claims across major platforms, with multi-website operations promoting identical low-cost merchandise under different brand names. A rule that names the templates and the non-existent shopfront, rather than only the false claim, gives enforcement a target earlier in the chain.

For buyers, the operational consequence is narrower and concrete. Campaign copy making strong performance claims, promotional pages using non-dismissible interstitials, creative assembled with generative tools, and credit, employment or housing campaigns using demographic parameters beyond the previously enumerated four all sit closer to a policy line on November 12 than they do now. Pinterest reserves the right to reject or remove any ad, and states that the rules may change at any time, language that both versions carry.

The timing sits against a platform under commercial pressure. Pinterest reported $1.18 billion in second-quarter 2026 revenue and 640 million monthly active users on August 4, 2026, with European advertising growth slowing to 12 percentas cross-border retailers retreated. A securities class action filed in March 2026 alleges the company concealed tariff-driven advertising declines. Tightening the categories that may be sold removes revenue at the margin, and the platform is doing it while building a self-serve tier aimed at advertisers with no campaign experience, the cohort least likely to read a guidelines diff before uploading creative.

Timeline

Summary

Who: Pinterest, Inc., and the advertisers, agencies and merchants operating accounts on its platform. The notice was sent by the Pinterest team to business account holders without a named spokesperson.

What: A bundled revision of the Terms of Service, Privacy Policy, Community Guidelines, Advertising Guidelines, Merchant Guidelines, GenAI Acceptable Use Guidelines, Advertising Services Agreement and the Ad Data Terms incorporated within it. The Advertising Guidelines gain a thirteen-point "Scams, deceptive goods and practices" section with five prohibitions that have no counterpart in the current text, shrink unacceptable business practices from thirteen entries to six under three new headings, add binary options to prohibited financial products, widen protected-group targeting restrictions from an enumerated list to a category, and narrow the landing page pop-up ban to non-dismissible pop-ups. The GenAI guidelines add account termination and feature loss as stated consequences of abuse. The Ad Data Terms were revised for connected TV buying and campaign extension across third-party publishers, and the Advertising Services Agreement now incorporates the Offsite Ad Terms by reference.

When: Announced September 4, 2026. Effective November 12, 2026. Both current and replacement policy pages are published in parallel until the effective date.

Where: Globally, across all Pinterest advertising products. Country-specific guidelines covering markets from Albania to Vietnam, and the trade sanctions list naming seven territories, carry over without substantive change.

Why: Pinterest describes the revision as making its terms easier to understand, clarifying dispute processes, clarifying policies on scams and deceptive goods, and reflecting changes to its advertising products. The document comparison shows added prohibited categories rather than clarification alone, and the contract changes align the paperwork with connected TV and offsite inventory acquired through the tvScientific transaction. Consent is passive: continued use after the effective date constitutes agreement.