Advertising Technology

Brand Safety: Strategies, Best Practices, and Platforms to Protect Your Ads

Brand safety is the set of policies, technologies, and practices that keep a brand’s advertising away from content that is illegal, dangerous, or universally unacceptable, such as terrorism, hate speech, or child exploitation. Brand suitability goes further, matching each ad to the content a specific brand considers appropriate for its values, audience, and tolerance for risk.

When I started in digital advertising, a media buyer could look at a list of sites and approve every placement. That world is gone. Programmatic platforms now place ads across millions of web pages, mobile apps, streaming channels, podcasts, and creator videos in the milliseconds it takes a page to load. Nobody reviews those placements by hand, which is why brand safety has become a discipline of its own, with standards bodies, accreditation programs, and an entire category of verification and intelligence platforms.

The stakes have shifted, too. The early concern was an ad for a family brand running beside an extremist video. Today, marketers also have to contend with made-for-advertising sites, mass-produced AI content, deepfakes, creator controversies, and blocklists so aggressive that they defund the trusted news environments advertisers say they value. This guide explains what brand safety is, why it matters, how the industry frameworks evolved, and which strategies, practices, and platforms help you protect your brand without starving your reach.

What Is Brand Safety?

Brand safety is the practice of preventing ads from appearing in, or funding, environments that would damage a brand’s reputation or that no responsible advertiser should support. It applies to every place an ad can run: websites, apps, connected TV, audio, social feeds, search results, influencer content, and, increasingly, AI-generated answers and chat experiences.

The industry has long organized brand safety around a common set of sensitive categories. The 4A’s Advertiser Protection Bureau published a brand suitability framework that groups content into categories such as:

  • Adult and explicit sexual content
  • Arms and ammunition
  • Crime, harmful acts, and human rights violations
  • Death, injury, and military conflict
  • Hate speech and acts of aggression
  • Illegal drugs, tobacco, vaping, and alcohol
  • Obscenity and profanity
  • Online piracy, spam, and harmful content
  • Terrorism
  • Sensitive social issues

For each category, the framework distinguishes between content that should never be monetized and content that carries high, medium, or low risk depending on how it is treated. A news report on a military conflict is very different from content that glorifies violence, even though both involve the same subject.

Brand Safety vs. Brand Suitability

Brand safety and brand suitability are related but not interchangeable:

  • Brand safety is the floor. It covers content that is illegal or universally unacceptable for advertising, regardless of the brand. Think terrorist propaganda, child sexual abuse material, or pirated content.
  • Brand suitability is everything above that floor. It reflects a brand’s own values, audience, and risk tolerance. A craft brewery might be comfortable next to nightlife and bar content that a children’s toy brand would avoid. A cybersecurity vendor might actively want placements beside coverage of data breaches.

Most modern programs treat safety as a non-negotiable exclusion and suitability as a set of adjustable tiers. The mistake many advertisers make is treating everything sensitive as a safety issue, which leads to overblocking.

Brand safety floor versus low, medium, and high risk suitability tiers

Why Brand Safety Is Critical

Brand safety matters because every impression is an implicit endorsement. When an ad funds a site or creator, the brand is financially supporting that content and, in the eyes of many consumers, associating itself with it. That creates risk in several ways:

  • Reputation and trust: A single screenshot of an ad beside offensive or fabricated content can become a social media story that outlasts the campaign.
  • Wasted spend: Unsafe environments often overlap with low-quality inventory, invalid traffic, and poor viewability, so the same controls that protect a brand also protect its budget.
  • Funding harm: Advertising revenue keeps disinformation networks, piracy sites, and content farms in business.
  • Accountability: Boards, regulators, and procurement teams increasingly expect marketers to show where their media dollars go.

Consumers notice. In its 2026 global study, DoubleVerify found that low-quality AI content is a growing source of brand risk:

Among 22,000 consumers surveyed across 22 markets, 42% said low-quality AI-generated ads would hurt their opinion of a brand, and 56% said they cannot consistently identify AI-generated content. Among 2,020 marketers surveyed, 53% said they are concerned about their ads appearing next to low-quality AI content.

DoubleVerify, Global Insights: Media Quality in the Age of AI, July 2026

Industry practitioners share the concern. Integral Ad Science (IAS) surveyed U.S. media experts for its annual industry outlook:

83% of U.S. media experts said AI-generated content on social media is a significant concern, 84% said third-party verification will be important for identifying AI-generated content on social platforms, and 69% cited ad content adjacency as a major challenge.

Integral Ad Science, 2026 IAS Industry Pulse Report, December 2025

Brand Safety Frameworks and Standards

Brand safety is not governed by law in most markets. Instead, it runs on voluntary industry frameworks, taxonomies, certifications, and audits. Understanding them helps you ask vendors and platforms the right questions.

GARM and Its Legacy

The World Federation of Advertisers (WFA) launched the Global Alliance for Responsible Media (GARM) in 2019 to create a shared vocabulary for advertisers, agencies, and platforms. GARM built on the 4A’s work with its Brand Safety Floor and Suitability Framework and later added an Adjacency Standards Framework to define how close an ad can run to sensitive content.

The WFA discontinued GARM in August 2024, shortly after X sued the WFA and several advertisers, alleging an illegal boycott. In July 2026, the two parties announced a settlement in which the WFA agreed not to form or restart GARM or a similar initiative. No single organization has replaced it. Instead, the GARM-era definitions remain available through the Brand Safety Institute framework library, and verification vendors and platforms continue to use the floor and risk-tier vocabulary in their controls and reporting.

IAB Tech Lab Content Taxonomy

The IAB Tech Lab maintains the Content Taxonomy, a standardized way to label what a page, app, or video is about. The Tech Lab encourages migration to the version 3 series and identifies brand safety and contextual targeting as two of its typical uses. When sellers and buyers use the same taxonomy, a suitability setting in a demand-side platform (DSP) means the same thing as the content signal passed by a supply-side platform (SSP).

TAG Brand Safety Certified

The Trustworthy Accountability Group (TAG) runs the Brand Safety Certified program for companies across the supply chain. Certified companies must appoint a compliance officer, train staff, use tools to avoid ad misplacement across their monetizable transactions, regularly review inclusion and exclusion lists, and screen for pirate mobile apps. Asking whether your agency, DSP, and SSP partners hold the seal is a simple way to raise the baseline.

MRC Accreditation

The Media Rating Council (MRC) audits and accredits measurement and verification services. In October 2025, it issued a policy on property-level ad verification representations to end confusion about what accredited brand safety means. The MRC distinguishes property-level verification, which classifies domains and apps using text and keywords, from content-level brand safety, which also analyzes images, video, and audio. Services without content-level accreditation should no longer position property-level filtering as brand safety.

Accreditation is also expanding to social video. The MRC’s first quarter 2026 accreditation update confirmed that Zefr received accreditation in January 2026 for content-level brand safety and suitability labeling and reporting on YouTube in-stream video. When you evaluate a vendor, ask which specific services, platforms, and environments its accreditation covers.

Key Brand Safety Strategies

No single tool solves brand safety. The most effective programs layer policy, supply decisions, pre-bid filtering, post-bid verification, and a response plan so that each layer catches what the others miss.

The layered brand safety stack: policy, supply, pre-bid, verification, and response controls working together

Replace Keyword Blocklists With Contextual Intelligence

Keyword blocklists were the first line of defense, and they remain the bluntest. A list that blocks words like “shot,” “fire,” or “attack” will block a vaccine story, a band review, and a sports recap along with genuinely harmful content. Contextual classification uses natural language processing (NLP), computer vision, and audio analysis to understand the meaning of the full page or video rather than individual words. I’ve covered this shift in more depth in our article on contextual targeting for brand safety.

The cost of keyword blocking is measurable. IAS analyzed Reuters inventory with the publisher:

54% of Reuters news URLs that IAS contextual analysis classified as brand suitable would have been blocked by a typical keyword blocklist. Even in the lifestyle section, 4.27% of suitable URLs would have been blocked, and those URLs represented 13.5% of lifestyle impressions. Only 0.25% of the impressions IAS measured were blocked by keyword-only rules.

Anthony Vargas, Keyword Blocking Demonetized More Than Half Of Reuters’ Brand-Safe Stories, AdExchanger, April 2026

Build Inclusion Lists and Maintain Exclusion Lists

Exclusion lists (blocklists) remove domains, apps, channels, and creators you never want to fund. Inclusion lists (allowlists) flip the model, limiting spend to environments you have vetted. Inclusion lists offer more control and usually better performance but can reduce scale, so many brands use them for high-risk campaigns or new channels and rely on exclusion lists plus contextual controls elsewhere. Either way, review lists on a schedule. Stale lists block publishers that have improved and miss new threats.

Avoid Made-for-Advertising Sites and AI Slop

Made-for-advertising (MFA) sites exist to arbitrage traffic, with heavy ad loads, auto-refreshing units, and little original content. They are usually not unsafe in the traditional sense, but they waste money and erode brand perception. The newest variant is AI slop: low-value, mass-produced content generated primarily by AI with little human input. The Association of National Advertisers (ANA) tracks both in its programmatic benchmarks:

MFA exposure rose to 1.1% of programmatic spend in the first quarter of 2026, up from a range of 0.4% to 0.6% throughout 2025, with AI slop identified as an emerging subtype of made-for-advertising inventory.

ANA, Programmatic Transparency Benchmark, Q1 2026, May 2026

An analysis of open web programmatic spend found that AI slop accounted for 1.3% to 2.4% of spend, that 88% of AI slop inventory was also classified as MFA, and that advertisers’ exposure ranged from 0.11% to 13.84% of spend. The analysis identified 11,552 AI slop domains.

ANA, TAG, and Fiducia, ANA/TAG/Fiducia Analysis Quantifies Level of “AI Slop” in Digital Advertising Supply Chain for First Time, August 2026

One finding stands out: AI slop inventory often looks fine on traditional quality metrics such as invalid traffic and viewability. You need content quality signals, not just fraud and viewability filters, to catch it.

Practice Supply Path Optimization

Supply path optimization (SPO) reduces the number of intermediaries between your DSP and the publisher. Fewer, more direct paths make it easier to know where your ads run, cut fees, and reduce exposure to resold or spoofed inventory. Private marketplace (PMP) deals and curated marketplaces add another layer of control. The ANA’s benchmarks show the payoff:

In the first quarter of 2026, higher-performing advertisers converted 54.0% of programmatic spend into qualified impressions that were fraud-free, measurable, viewable, and free of MFA inventory, compared with 32.1% for lower performers, and they operated with more concentrated supply footprints.

ANA, Programmatic Transparency Benchmark, Q1 2026, May 2026

Extend Protection to Connected TV

Connected TV (CTV) used to be considered inherently safe because it resembled television. In reality, CTV bid requests often lack reliable information about which show or channel is playing, and the channel attracts fraud schemes such as device spoofing. Look for partners that can identify the actual program and app in the bid request, apply CTV-specific fraud detection, and report placements at the show level.

CTV accounted for 45.6% of programmatic spend in the third quarter of 2025, and 99.1% of tracked programmatic spend ran in environments classified as low risk under the benchmark’s first brand safety and suitability measurements.

ANA, Programmatic Transparency Benchmark, Q3 2025, November 2025

Protect Social, Creator, and UGC Placements

Social feeds and user-generated content (UGC) change by the second, and most platforms are walled gardens where buyers cannot deploy their own tags freely. Use each platform’s native suitability controls, such as inventory filters on Meta and TikTok, and add third-party verification where the platform supports integrations. For influencer and creator partnerships, vet the creator’s history, audience, and past controversies before signing, and monitor content throughout the campaign.

87% of U.S. media experts said the brand safety and suitability of the influencer or creator they partner with is essential, and 82% said creator suitability is as important as the suitability of the content their ads appear next to.

Integral Ad Science, 2026 IAS Industry Pulse Report, December 2025

Prepare for Generative AI and Deepfakes

Generative AI cuts both ways. It powers the classification models that make contextual brand safety possible, and it also produces synthetic content, fabricated news, and deepfakes that can impersonate executives or misuse a brand’s identity. Brands should monitor for unauthorized use of their logos, spokespeople, and likenesses, ask verification partners how they detect AI-generated content, and set clear policies for when their own ads can appear in AI chat and search experiences. We explored the broader risk in our article on deepfake technology in marketing.

Stop Overblocking News

News environments are among the most trusted, engaged audiences a brand can reach, yet blunt brand safety settings routinely defund them. Research from ThinkNewsBrands in Australia shows how much is at stake:

An analysis of 1,200 articles from 10 news publishers found that 57% contained at least one exclusion trigger, and 38% of low-sensitivity articles that were suitable for advertising were incorrectly excluded. The study estimated that about A$180 million in Australian news advertising spend is at risk from overly broad brand safety settings.

ThinkNewsBrands, Beyond Brand Safety, September 2026

The fix is to separate genuinely harmful content from serious journalism. Use contextual suitability tiers instead of keyword lists for news, allow low-risk coverage of sensitive topics, and review what your settings are blocking, not just what they are letting through.

Brand Safety Best Practices

Strategy sets direction. These practices make it stick across teams, agencies, and platforms:

  • Write a brand suitability policy. Document your floor, your acceptable risk tier for each sensitive category, and any brand-specific exceptions. Share it with every agency and platform partner so settings are consistent.
  • Calibrate by campaign. A product launch, a crisis-sensitive period, and an always-on performance campaign may warrant different suitability settings.
  • Prefer context over keywords. Keep keyword lists short, targeted, and reviewed. Rely on contextual classification for nuance.
  • Combine pre-bid and post-bid controls. Pre-bid filters stop bad impressions before you pay. Post-bid verification confirms where ads actually ran and catches what the filters missed.
  • Measure fraud and quality alongside safety. Track invalid traffic (IVT), viewability, MFA exposure, and AI slop. Our guide to ad fraud covers the fraud side in detail.
  • Demand transparency. Require log-level or placement-level reporting from DSPs, SSPs, and platforms, and ask which certifications and accreditations each partner holds.
  • Review lists and settings quarterly. Refresh inclusion and exclusion lists, audit blocked inventory, and remove outdated keywords.
  • Monitor the conversation. Social listening and media monitoring alert you when a placement, creator, or AI-generated impersonation starts drawing attention.
  • Have a response plan. Decide in advance who can pause campaigns, how you will investigate an incident, and how you will communicate with customers and partners.
  • Balance safety with reach. Track how much inventory your settings block and the performance of the environments you include. Overblocking is a cost, not a safe default.

Brand Safety Monitoring Platforms

Brand safety platforms fall into three broad groups: ad verification and media quality platforms that measure where ads run, contextual and fraud specialists that filter inventory before you bid, and listening and reputation tools that monitor what is being said about your brand. Most enterprise advertisers use at least one from each group.

PlatformFocusWhat It Does
DoubleVerifyAd verification and media qualityAn AI-driven media effectiveness platform that measures brand suitability, fraud, and viewability across the open web, CTV, social, and commerce media, including verification for AI-generated content.
Integral Ad ScienceAd verification and media qualityMeasures and optimizes media quality, including brand safety and suitability, contextual targeting, fraud, and viewability across CTV, social, and the open web, with controls for low-quality generative AI content.
ZefrSocial and video suitabilityUses AI content classification to measure and target brand suitable video and social feed content on YouTube, Meta, and TikTok, with solutions for AI-generated content and brand misappropriation.
Channel FactoryVideo and CTV suitabilityBuilds brand suitable inclusion lists and contextual video campaigns across YouTube, Meta, TikTok, and CTV, focused on aligning ads with values and performance goals.
PixalateFraud, privacy, and complianceProvides fraud protection, privacy, and compliance analytics for CTV, mobile apps, and websites, including app and channel ratings that support brand safety decisions.
HUMANBot and fraud preventionProtects advertising from bots and sophisticated invalid traffic from the bid request through the click, helping keep spend on real human audiences.
GumGumContextual intelligenceAnalyzes the full content of a page or video, including text, images, video, and audio, to deliver contextual targeting and brand suitability signals before every bid. Previously marketed as Verity.
Peer39Pre-bid contextual dataSupplies pre-bid contextual, suitability, quality, fraud, and MFA signals across CTV, audio, social, and the web, available directly in major DSPs.
NewsGuardNews reliability ratingsRates the reliability of news and information sources so advertisers can support credible journalism while avoiding misinformation sites.
Ad Fontes MediaNews bias and reliabilityRates news sources for bias and reliability, best known for its Media Bias Chart, giving advertisers a way to evaluate news inventory instead of blocking it.
BrandwatchSocial listeningA consumer intelligence and social listening suite that tracks brand mentions, sentiment, and emerging crises, with AI-powered alerts.
MeltwaterMedia and social monitoringCombines media monitoring, social listening, and AI signal tracking to show how a brand is covered and discussed across news and social channels.
SprinklrCustomer experience and listeningAn AI-native customer experience platform whose consumer intelligence capabilities monitor real-time conversations to surface risks and opportunities.

One name you may still see in older brand safety guides is Moat. Oracle retired Moat along with the rest of its advertising business, and the Oracle Advertising end-of-life FAQ confirms those services ended on September 30, 2024. If your measurement stack or vendor contracts still reference Moat, it is time to migrate.

How to Choose a Platform

Start with where your money goes. If most of your budget runs through programmatic open web and CTV, a verification platform paired with pre-bid contextual data is the foundation. If social and creators dominate, prioritize partners that integrate with those platforms and hold relevant accreditation. If reputation risk is your primary concern, add listening and monitoring. In every case, ask for:

  • The specific MRC accreditations held, by service and environment
  • Whether classification covers text only or also images, video, and audio
  • Coverage for CTV, social platforms, creators, and AI-generated content
  • Pre-bid integration with your DSPs and post-bid reporting granularity
  • Visibility into what your settings block, not just what they allow

If you’re evaluating verification partners, DoubleVerify is a good place to start. You can read our earlier look at the DoubleVerify platform as well.

Explore DoubleVerify

Frequently Asked Questions

What is the difference between brand safety and brand suitability?

Brand safety is the floor: content that is illegal or unacceptable for any advertiser, such as terrorism or child exploitation. Brand suitability sits above that floor and reflects each brand’s own values and risk tolerance, so content that is suitable for one brand may be avoided by another.

What replaced GARM?

No single organization replaced GARM. The World Federation of Advertisers discontinued it in August 2024 and agreed in a July 2026 settlement with X not to restart it or a similar initiative. Its brand safety floor and suitability definitions remain in use through the Brand Safety Institute framework library and in the controls offered by verification vendors and platforms.

Are keyword blocklists still useful?

Short, targeted keyword lists can still help for specific brand concerns, but broad lists overblock legitimate content, especially news. Contextual classification that analyzes the meaning of the full page or video is more accurate and protects reach.

What is AI slop in advertising?

AI slop is low-value, mass-produced content generated primarily by AI with little human input. Much of it overlaps with made-for-advertising sites, and it can pass traditional fraud and viewability checks, so advertisers need content quality signals to avoid it.

How do I protect my brand on CTV?

Work with partners that can identify the actual app and program in CTV bid requests, apply CTV-specific fraud detection, and report placements at the show or channel level. Supply path optimization and direct or private marketplace deals add further control.

Which brand safety platform should I use?

Most advertisers combine an ad verification platform such as DoubleVerify or Integral Ad Science, a pre-bid contextual or fraud specialist, and a social listening tool. Choose based on where your budget runs, the accreditations each vendor holds, and the environments it covers.

Related Articles