What GARM was
The Global Alliance for Responsible Media was an initiative of the World Federation of Advertisers. For brand safety and suitability it became a shared reference point: a common set of categories for content that advertisers would avoid entirely, and for content they would accept at different levels of risk. Industry documents adopted its vocabulary. The IAB’s 2023 podcast buyer and seller checklist, for example, refers to the “GARM brand safety & suitability floor” next to the IAB Content Taxonomy (IAB, 2023-10).
Why it ended
The WFA discontinued GARM on August 9, 2024. In its statement, the WFA said that allegations made against the initiative had “significantly drained its resources and finances” (WFA, retrieved 2026-10-05).
The 2026 settlement
On July 29, 2026, the WFA and X Corp. settled the litigation over GARM. As part of its announcement, the WFA stated that it “will not form or restart GARM or a similar initiative” (WFA, 2026-07-29). For buyers, that closes the question of whether a shared industry framework will return under the WFA. It will not.
Where its categories still appear
The framework outlived the organization. Its categories still appear in vendor tools and in older industry documents, such as the IAB checklist above. That is not a problem in itself, as long as a brand knows what it is relying on. When a vendor scores content “to GARM”, the brand is relying on that vendor’s reading of a framework that nobody now maintains.
The plain way to describe this in a media plan is to name the actual framework: the brand’s own policy, or the scoring vendor’s categories by name. Calling GARM a current standard is inaccurate.
How suitability works in podcasts now
Podcast brand safety now runs on transcripts. According to the IAB’s 2023 checklist, it depends on AI review of episode transcripts, and buyers should define their thresholds for suitability and a floor for safety, then select a vendor to score against them (IAB, 2023-10). Most publishers already use these tools: the IAB reported that 69% of podcast publishers used brand safety tools and 62% used suitability tools, on 2023 data (IAB, 2023-10-26). Barometer is one vendor that scores episodes from transcripts, and trade press reports Acast among its clients (AdExchanger, retrieved 2026-10-05).
Those scores are a starting point. A transcript score catches language and topics; it does not catch a host’s tone, a recurring segment the brand would dislike, or a past sponsor read that made a claim the brand could not make. A human review of the back catalog still has to sit on top.
What to ask a scoring vendor
A transcript score is only as useful as the rules behind it, and those rules now differ from vendor to vendor. Before relying on one, ask it to state in writing:
- Which framework its categories follow, and who maintains that framework now.
- Whether it scores full transcripts or samples, and how it handles video podcasts, where the picture can matter as much as the words.
- Whether a score describes the whole episode or the part around the ad break.
- How soon a new episode is scored after release, which matters for shows booked a few weeks ahead.
- What each score level means in plain terms, and which passages in the transcript produced it.
The answers belong in the brand’s policy, so that a vetting memo can say which vendor scored a show, against which categories, and what the score was based on.
How it works on YouTube
For paid ads on YouTube, Google Ads offers three inventory types (Expanded, Standard, the default, and Limited) and lets advertisers exclude content such as live streams (Google Ads Help, retrieved 2026-10-05). Those settings govern where paid ads run. A sponsored integration sits inside a creator’s own video, so its suitability is decided when the channel is chosen, before the booking.
Contents of a brand policy
With no shared framework, every brand needs a short written policy. It should cover six things:
- The floor: categories of content the brand will never appear next to.
- The thresholds: categories the brand accepts with a condition, and what the condition is.
- The method: the scoring vendor and its framework, named, plus the human review on top.
- Who decides exceptions, and how quickly.
- How often a booked channel is re-checked during a long flight.
- The procedure for a live placement in breach: who is notified, and whether the placement is withdrawn or made good.
A policy like this can fit on a page or two. Its value is that the vetting memo for every channel can cite it, so the reason for each booking, condition or decline is written down.