Harmful-Content Enforcement Surge 2026: Brand Safety

Quick Answer

In May 2026, total platform content-moderation decisions reported to the EU DSA Transparency Database fell 17.8% month-over-month, yet enforcement against the most harmful categories surged: self-harm enforcement rose 50% (1.77M to 2.65M decisions), violence rose 30% (3.78M to 4.90M), and cyber-violence rose 69% (613K to 1.03M). Illegal or harmful speech rose 16% and consumer-information enforcement rose 22%. Across all categories, 93.6% of decisions were automated. The headline is a divergence: platforms eased up on routine moderation while cracking down hardest on the content most damaging to people and to brands. For advertisers this rewrites the brand-safety map. A falling total enforcement volume does not mean a safer environment — it means moderation effort is being concentrated on severe harm, which shifts where adjacency risk sits. Branded ads are less likely to be pulled into routine policy churn but operate alongside an automated enforcement system aimed squarely at violent, self-harm, and cyber-violence content, where both the harm and the over-removal risk are highest. The defensible response is to stop treating brand safety as a single platform setting and to verify inventory and adjacency controls, category and keyword exclusions, sensitivity tiers, and third-party verification against the 2026 environment. Figures are drawn from the EU DSA Transparency Database (CC BY 4.0). Read the full breakdown on the Platform Enforcement Index for May 2026 and track changes on the Policy Change Tracker.

Predict ad disapprovals before you submit — try the free Meta Rejection Predictor

Why Harmful-Content Enforcement Surged While Totals Fell

In May 2026, the total volume of platform content-moderation decisions reported to the EU DSA Transparency Database fell 17.8% from the prior month. Read alone, that looks like platforms easing up. But underneath the falling total, enforcement against the most harmful categories surged: self-harm enforcement rose 50%, violence rose 30%, and cyber-violence rose 69%. Across everything, 93.6% of decisions were automated.

The story is a divergence, not a loosening. Platforms eased routine moderation while concentrating effort on the content most damaging to people — and to brands. For advertisers, this rewrites the brand-safety map. A falling total enforcement volume does not mean a safer environment; it means moderation effort has been redistributed toward severe harm, which changes where adjacency risk sits and how predictable it is.

"Total enforcement fell 17.8% in May 2026, yet self-harm (+50%), violence (+30%) and cyber-violence (+69%) enforcement all surged — 93.6% of it automated.
— Source: EU DSA Transparency Database (CC BY 4.0), May 2026"

This guide breaks down the data, explains the divergence, and translates it into a brand-safety playbook: what surging harm enforcement means for adjacency risk, which controls match the 2026 environment, and which verticals carry the most exposure. For the full breakdown see the Platform Enforcement Index for May 2026.

The 2026 Enforcement Data: What Surged, What Eased

The category movers tell the story. While the overall total fell, the severe-harm categories rose sharply.

Category Movers, April to May 2026

Category	April	May	Change
Cyber Violence	613,043	1,034,222	+69%
Self-Harm	1,770,774	2,649,075	+50%
Violence	3,782,207	4,898,724	+30%
Consumer Information	722,049	883,250	+22%
Illegal / Harmful Speech	4,998,267	5,778,406	+16%

Against those increases, the total fell to about 102.4 million decisions, down 17.8% month-over-month, with roughly 43.4 million pieces of content removed and about 93.6% of all decisions made by automated means, based on our analysis of the EU DSA Transparency Database (CC BY 4.0), which platforms populate through mandatory statements of reasons.

How to Read the Numbers

Two caveats matter. The database is self-reported, so platforms categorise inconsistently, and figures reflect moderation decisions rather than a perfect measure of harm prevalence. A monthly total can also swing on a single large coordinated-behaviour sweep. The signal is the concentration on severe harm, not the precise totals. For the EU framework see the European Union DSA compliance guide.

The Divergence: Automation Aimed at the Worst Content

Why would platforms ease total moderation but enforce harder on harmful content? Four dynamics combine.

The Drivers

Automation tuning: With 93.6% of decisions automated, the enforcement mix is driven by how detection thresholds are configured — and platforms tune toward the categories with the greatest legal and reputational exposure.
Regulatory pressure: The EU DSA, Australia's and the UK's Online Safety Acts, and the US TAKE IT DOWN Act all concentrate obligations and penalties on severe harm.
Volume volatility: Total counts swing month to month for reasons unrelated to safety posture — a one-off spam sweep can inflate one month and deflate the next.
Resource reallocation: As trust-and-safety teams have restructured, automated systems carry more load, steered toward the highest-priority harms.

The net effect is concentration, not a blanket loosening. The environment is neither uniformly safer nor more dangerous — it is more concentrated, with severe-harm content enforced harder and routine content enforced less. That redistributes adjacency risk rather than eliminating it.

What Surging Harm Enforcement Means for Brand Safety

The redistribution of enforcement cuts both ways for advertisers, and the net is a less predictable adjacency environment.

Two Directions of Effect

Protective: Heavier enforcement of self-harm, violence, and cyber-violence content removes more of what a brand would never want to appear beside.
Collateral: The same automated systems sweep up borderline or contextual content, remove and reinstate on appeal, and create volatile moderation states on specific surfaces — making adjacency less predictable.

The key consequence is that a brand cannot infer its adjacency risk from headline enforcement totals. A falling total can coexist with rising risk in the surfaces and categories a campaign actually touches. Brand safety must therefore be configured and verified at the campaign level, not inferred from platform-wide numbers. The reputational stakes are asymmetric — the cost of a single high-profile adjacency to violent or self-harm content vastly exceeds the cost of conservative settings — so most brands should err toward stricter controls. To audit creative and placements use the AI Compliance Audit.

Adjacency Risk in a Shifting Moderation Landscape

Adjacency risk — the chance branded ads appear beside content a brand would never sponsor — is the central brand-safety concern, and the 2026 data makes it harder to predict.

What Makes 2026 Different

Factor	Effect on adjacency risk
93.6% automation	Faster removal but more collateral over-removal and reinstatement churn
Concentration on severe harm	Risk shifts toward the most reputationally damaging categories
Weaker industry coordination	After GARM wound down its activities in 2024, advertisers rely more on platform and third-party controls
Monthly volatility	The moderation state of a surface can change between campaign flights

The implication is that adjacency cannot be set once and forgotten. It must be configured conservatively, verified independently, and monitored continuously. To tune exclusion language use the Keyword Risk Checker and to follow platform-policy shifts use the Policy Change Tracker.

Brand-Safety Controls That Match the 2026 Environment

No single control is sufficient when enforcement is heavily automated and concentrated on severe harm. The defensible approach is a layered, documented stack.

The Control Stack

Inventory and adjacency settings: Control where ads appear and choose the most conservative adjacency consistent with campaign goals.
Content-category exclusions: Align exclusions to the surging-harm categories — violence, self-harm, cyber-violence, harmful speech — not a generic template.
Keyword and topic exclusions: Tune precisely; overly broad blocks suppress legitimate news and cut reach.
Sensitivity tiers: Apply the tier matching your risk tolerance, confirmed at campaign and account level.
Third-party verification: Independent measurement is more credible than platform-supplied adjacency reporting.
Continuous monitoring: Review configuration on a fixed cadence and after any major policy change.

Documentation underpins all of it: keep a dated record of controls per campaign so any incident can be answered with evidence. To audit placements use the AI Compliance Audit.

High-Risk Verticals and Sensitive Categories

Risk is highest where surging-harm adjacency, sensitive audiences, and regulated verticals overlap.

The Exposure Map

Adjacency to surging-harm categories: Campaigns near violent, self-harm, or cyber-violence content carry both higher catch rates and higher reputational severity.
Audience sensitivity: Campaigns that can reach minors face additional children's-safety scrutiny; adjacency in a youth context is among the most damaging failures.
Regulated verticals: Healthcare, finance and crypto, gambling, and alcohol carry layered verification, disclosure, licensing, and placement requirements on top of general adjacency risk.
Own-content risk: Brands using sensitive imagery, strong language, or AI-generated creative can have their own content flagged by automated enforcement.

Map the brand's exposure across these axes and apply the strictest settings where they intersect. For industry-specific obligations see the healthcare social-media compliance guide or the financial services ad compliance guide.

2026 Brand-Safety Checklist

[ ] Brand risk tolerance defined and documented per account
[ ] Inventory and adjacency settings reviewed and set conservatively on every campaign
[ ] Content-category exclusions aligned to surging-harm categories (violence, self-harm, cyber-violence)
[ ] Keyword and topic exclusions tuned precisely to the brand profile
[ ] Sensitivity / content-control tier confirmed at campaign and account level
[ ] Third-party brand-safety verification engaged where available
[ ] Own creative audited for content that could be flagged by automated enforcement
[ ] Regulated-vertical obligations (healthcare, finance, gambling, alcohol) mapped where relevant
[ ] Brand-safety configuration documented with dates per campaign
[ ] Continuous monitoring cadence set; Enforcement Index and Policy Change Tracker reviewed

For multi-jurisdiction stress-testing use the Legal Compliance Scan and for the current data see the Platform Enforcement Index.

Frequently Asked Questions

What does the 2026 enforcement data actually show, and where does it come from?

The 2026 enforcement data shows a clear divergence — total platform content-moderation decisions fell 17.8% month-over-month in May 2026 while enforcement against the most harmful categories surged — and the figures are drawn from the European Union's Digital Services Act Transparency Database, which platforms are legally required to populate with a statement of reasons for each moderation decision. The specific movements are instructive. Self-harm enforcement rose 50%, from about 1.77 million decisions to about 2.65 million. Violence enforcement rose 30%, from about 3.78 million to about 4.90 million. Cyber-violence enforcement rose 69%, from about 613,000 to about 1.03 million — the steepest proportional increase. Illegal or harmful speech rose 16%, and consumer-information enforcement rose 22%. Against those increases, the overall volume of decisions fell to about 102.4 million, down 17.8% from the prior month, and 93.6% of all decisions were automated. The data source matters for credibility: the DSA Transparency Database is an official EU dataset published under a Creative Commons Attribution 4.0 licence, populated by the platforms themselves through mandatory statements of reasons, which makes it the most comprehensive public record of platform moderation in existence — though it carries known caveats, including that platforms self-report and categorise inconsistently, and that figures reflect decisions rather than a perfect measure of harm prevalence. For advertisers the takeaway is that a falling total is not a signal of a safer environment; it is a signal that moderation effort is being concentrated on severe harm. The full month-by-month breakdown, with platform-level and action-type detail, is published on the Platform Enforcement Index for May 2026, and ongoing platform-policy movements are tracked on the Policy Change Tracker. A few additional features of the dataset help an advertiser read it correctly rather than over-interpret single numbers. Enforcement under the DSA is applied overwhelmingly EU-wide, so the figures reflect decisions with broad territorial scope rather than country-by-country moderation, which is why a per-country ranking is not meaningful. Some movements are large in percentage terms but small in absolute terms — a category or platform can show a triple-digit percentage increase off a small base while contributing little to the overall picture — so proportional changes should always be read alongside absolute volumes. And because the figures are statements of reasons filed by the platforms themselves, they measure moderation actions taken, not the underlying prevalence of harmful content; a rise in enforcement can reflect better detection rather than more harm, and a fall can reflect the absence of a one-off sweep rather than more harm slipping through. For the advertiser, the durable signal is not any single month's total but the direction: a sustained concentration of enforcement on severe-harm categories. The organizing principle is that the data is official, self-reported, and shows concentration of enforcement on the worst content even as routine moderation eased.

Why would platforms ease total moderation but increase enforcement on harmful content?

Platforms eased total moderation while increasing enforcement on the most harmful content because automated moderation systems are increasingly tuned to prioritise severe-harm categories, because regulatory and reputational pressure is concentrated on those categories, and because routine, lower-severity moderation is both less risky to deprioritise and more prone to volume fluctuation — and the net effect is a deliberate concentration of effort rather than a uniform pullback. Several dynamics combine. First, automation tuning: with 93.6% of decisions automated, the mix of what gets enforced is driven heavily by how detection models and thresholds are configured, and platforms have strong incentives to tune those systems toward the categories that carry the greatest legal and reputational exposure — self-harm, violence, child safety, and cyber-violence. Second, regulatory pressure: the EU's Digital Services Act, Australia's Online Safety Act, the UK's Online Safety Act, and the US TAKE IT DOWN Act all concentrate obligations and penalties on severe harm, so platforms rationally direct enforcement capacity there. Third, volume volatility: total decision counts can swing month to month for reasons unrelated to safety posture — a single large spam or coordinated-behaviour sweep can inflate one month's total and deflate the next — so a falling total can reflect the absence of a one-off sweep rather than a policy of leniency. Fourth, resource reallocation: as platforms have restructured trust-and-safety teams, automated systems carry more of the load, and that load is steered toward the highest-priority harms. For advertisers the why matters because it shapes the risk: the environment is not uniformly safer or more dangerous, it is more concentrated, with severe-harm content being enforced harder and routine content being enforced less. This means adjacency risk is being redistributed, not eliminated. To understand how platform policies are themselves changing see the Policy Change Tracker and for the EU framework that drives much of this see the European Union DSA compliance guide. The content-type mix in the data reinforces why this concentration matters for advertisers specifically. Video dominates enforcement volume, accounting for the largest share of actioned content, with text, image, and a meaningful and growing slice of synthetic media following — which maps directly onto the formats brands advertise in and around. As automated systems concentrate on severe-harm video and synthetic content, the surfaces where short-form video advertising lives become both more actively policed and more volatile, because that is where enforcement effort is going. There is also a feedback loop worth noting: regulatory regimes that attach the heaviest penalties to severe harm give platforms a rational reason to over-invest detection capacity there, even at the expense of consistency on lower-severity categories, which is part of why routine moderation can ease while severe-harm enforcement climbs. None of this implies a platform is becoming safer or more dangerous overall; it implies the risk has a new shape, concentrated in specific categories and formats. The accurate framing is concentration driven by automation tuning and regulatory pressure, not a blanket loosening.

How does surging harm enforcement change brand-safety risk for advertisers?

Surging harm enforcement changes brand-safety risk by redistributing where adjacency exposure sits rather than reducing it overall, and advertisers should respond by treating brand safety as a configured, verified system rather than a platform default, because the 2026 environment concentrates both genuine harm and automated over-removal in the categories most damaging to brands. The redistribution works in two directions. On one hand, heavier enforcement of self-harm, violence, and cyber-violence content means platforms are removing more of the content a brand would never want to appear beside, which is protective. On the other hand, the same automated systems that drive a 50% rise in self-harm enforcement and a 69% rise in cyber-violence enforcement also generate collateral effects: borderline or contextual content can be swept up, content can be removed and reinstated on appeal, and the moderation state of a given surface can be volatile, which makes the adjacency environment less predictable. A brand cannot infer its adjacency risk from headline enforcement totals, because a falling total can coexist with rising risk in the specific surfaces and categories a campaign touches. The practical consequence is that brand safety must be configured and verified at the campaign level: inventory and adjacency settings, content-category exclusions aligned to the harm categories most damaging to the brand, keyword and topic exclusions tuned to the brand's profile, sensitivity tiers, and independent third-party verification. It also means monitoring is continuous rather than one-off, because the moderation environment shifts month to month. The reputational stakes are asymmetric: the cost of a single high-profile adjacency to violent or self-harm content vastly exceeds the cost of conservative settings, so the rational posture for most brands is to err toward stricter controls. To audit creative and placements use the AI Compliance Audit, to tune exclusion language use the Keyword Risk Checker, and to follow the data see the Platform Enforcement Index. The reinstatement dynamic is the part advertisers most often overlook, and it is a direct product of high automation. When automated systems remove content quickly, a meaningful share is appealed and reinstated under DSA complaint rights, which means the moderation state of a given surface is not static — content disappears and reappears, and the adjacency profile of a placement can shift within a single campaign flight. For a brand, this defeats any model that treats a surface as permanently safe or unsafe based on a one-time assessment. It also means that contextual content — borderline material that is neither clearly violating nor clearly benign — is precisely the category where automated decisions are least reliable and most volatile, and where a brand's exposure is hardest to predict. The defensive implication is that brand-safety configuration cannot be a launch-day task that is then forgotten; it has to be monitored continuously and verified independently, because the environment it is protecting against is itself changing week to week. The organizing principle is that risk is redistributed and unpredictable, so brand safety must be configured, verified, and continuously monitored.

What brand-safety controls best match the 2026 enforcement environment?

The brand-safety controls that best match the 2026 enforcement environment are a layered stack — inventory and adjacency settings, content-category exclusions, keyword and topic exclusions, sensitivity tiers, third-party verification, and continuous monitoring — configured conservatively at the campaign level and documented, because no single control is sufficient when enforcement is heavily automated and concentrated on severe harm. Start with inventory and adjacency: control where ads can appear and choose the most conservative adjacency consistent with campaign goals, since defaults frequently favour reach over safety. Layer content-category exclusions aligned to the specific harm categories the data shows are surging — violence, self-harm, cyber-violence, and harmful speech — so that the brand's exclusions track the actual risk environment rather than a generic template. Add keyword and topic exclusions tuned to the brand's profile, recognising that overly broad keyword blocks can also suppress legitimate news and reduce reach, so the list should be precise. Use sensitivity or content-control tiers where platforms offer them, and confirm they are applied at both campaign and account level. Engage third-party brand-safety and verification partners, because independent measurement is more credible than platform-supplied adjacency reporting — a lesson reinforced by the broader scrutiny of platform transparency in 2026. Finally, treat monitoring as continuous: the enforcement environment shifts month to month, so brand-safety configuration should be reviewed on a fixed cadence and after any major platform-policy change. Documentation underpins all of it — keep a dated record of the controls applied per campaign so that any adjacency incident can be answered with evidence of the safeguards in place. The reputational asymmetry justifies conservatism: the downside of a single severe adjacency dwarfs the cost of stricter settings. To audit placements use the AI Compliance Audit and to monitor platform-policy shifts use the Policy Change Tracker. One trade-off deserves explicit attention because it is where brand-safety configuration most often goes wrong: keyword and topic exclusions can be set so broadly that they suppress legitimate, high-value inventory. Aggressive block lists that catch news, current-affairs, and contextual content can starve a campaign of reach and push delivery toward lower-quality inventory, which is a real cost rather than a free safety margin. The answer is precision rather than breadth — exclusions tuned to the brand's actual risk categories, tested against delivery, and refined over time — paired with the higher-leverage controls of inventory adjacency settings and sensitivity tiers, which reduce risk without the collateral reach loss of an over-broad keyword list. Independent verification then closes the loop by measuring what actually ran, so the configuration can be adjusted on evidence rather than assumption. The point is that conservative does not mean blunt: the most effective 2026 configurations are conservative in posture but precise in execution, and they are revisited as both the enforcement environment and campaign delivery data evolve. The practical summary is a documented, conservative, layered configuration reviewed continuously.

Which industries and content categories carry the highest brand-safety risk in 2026?

In 2026 the highest brand-safety risk concentrates where surging harm enforcement meets sensitive audiences and regulated verticals — campaigns near violent, self-harm, and cyber-violence content, campaigns reaching minors, and advertisers in regulated industries such as healthcare, finance, gambling, and alcohol — and the risk is compounded for brands whose own creative or targeting touches sensitive topics. The category data points directly at the surfaces to avoid: with violence enforcement up 30%, self-harm up 50%, and cyber-violence up 69%, the content most damaging to brands is also the content platforms are enforcing hardest, which makes adjacency to those categories both more likely to be caught and more reputationally severe if missed. Audience sensitivity multiplies the risk: campaigns that can reach minors face additional scrutiny under children's-safety regimes, and adjacency to harmful content in a youth context is among the most damaging brand-safety failures. Regulated verticals carry layered exposure: healthcare and supplement advertisers face claim-substantiation and platform restriction risk, financial-services and crypto advertisers face verification and disclosure requirements, gambling advertisers face licensing and certification gates, and alcohol advertisers face age-gating and placement restrictions — and all of these sit on top of the general adjacency risk. Brands whose creative uses sensitive imagery, strong language, or AI-generated content face an additional layer, because automated enforcement can flag the brand's own content, not only the content around it. The defensible approach is to map the brand's specific exposure across three axes — adjacency to surging-harm categories, audience sensitivity, and vertical-specific regulation — and to configure controls accordingly, applying the strictest settings where the axes intersect. For industry-specific obligations see the relevant guide, such as the healthcare social-media compliance guide or the financial services ad compliance guide, and to track regulated-vertical policy changes use the Policy Change Tracker. It helps to make the vertical exposures concrete rather than abstract. A healthcare or supplements brand carries claim-substantiation risk on its own creative plus adjacency risk around self-harm content, a combination that is especially fraught given the surge in self-harm enforcement. A financial-services or crypto advertiser carries verification and disclosure obligations that automated systems enforce aggressively, so its own ads are more likely to be flagged even before adjacency is considered. A gambling advertiser operates behind licensing and certification gates and cannot afford an account-level strike, making conservative settings a commercial necessity rather than a preference. An alcohol brand faces age-gating and placement restrictions that interact with audience-sensitivity risk. And any brand that can reach minors inherits children's-safety scrutiny on top of everything else. Mapping exposure across the three axes — adjacency to surging-harm categories, audience sensitivity, and vertical-specific regulation — lets a brand apply the strictest controls precisely where the axes intersect, rather than applying uniform settings that are simultaneously too loose for high-risk campaigns and too restrictive for low-risk ones. The organizing principle is that risk is highest where surging-harm adjacency, sensitive audiences, and regulated verticals overlap.

Should advertisers reduce spend on platforms with high harmful-content enforcement?

High harmful-content enforcement is not, by itself, a reason to reduce spend — heavy enforcement of severe-harm categories is protective, and the right response is to match controls and verification to the environment rather than to retreat from it — but the decision should be governed by the brand's risk tolerance, the specific surfaces a campaign touches, and independent measurement rather than headline enforcement figures. It is a mistake to read a high enforcement count as a signal of a dangerous platform: a platform removing large volumes of violent or self-harm content is demonstrating active moderation of exactly the content brands fear, so high enforcement can be reassuring rather than alarming. It is equally a mistake to read a falling total as a signal of safety, because the total can fall while risk rises in specific surfaces. The decision should therefore rest on three inputs. First, risk tolerance: a children's brand, a regulated-industry advertiser, or a public-sector account will weight adjacency risk far more heavily than a B2B brand, and should configure and possibly restrict spend accordingly. Second, surface specificity: assess where the campaign actually runs and what content it is adjacent to, using independent verification rather than platform-supplied adjacency reporting. Third, measurement: run bounded tests with conservative settings, measure adjacency and brand outcomes independently, and review on a fixed cadence. For most advertisers the outcome will be continued spend with tighter, documented controls and continuous monitoring; for the highest-sensitivity brands, a targeted reduction on specific surfaces or formats may be warranted. The decision should be documented so it is defensible to internal stakeholders, and it should be revisited as the enforcement environment shifts month to month. To inform the decision with current data see the Platform Enforcement Index and to audit placements use the AI Compliance Audit. The deciding discipline is to separate the question of whether the platform is moderating severe harm from the question of whether a specific campaign is safe on it — because they have different answers and require different evidence. High severe-harm enforcement answers the first question reassuringly, but it says little about the second, which depends on the surfaces a campaign actually touches, its audience, and the controls applied. An advertiser that conflates the two will either wrongly flee a platform that is actively policing the content it fears, or wrongly relax on a platform whose aggregate numbers look fine while the campaign's specific placements carry risk. The way to keep the questions distinct is independent, surface-level measurement: verify where the ads ran and what they ran beside, and let that evidence — not the platform-wide enforcement headline — drive the spend decision. Reviewed on a fixed cadence, this turns brand safety from a reactive headline response into a managed, evidence-based program. The organizing principle is proportionality: match controls and verification to risk tolerance and surface specificity, and decide on measured evidence rather than headline counts.

Don't miss the next policy change.

Create a free account — track every policy change across 8 platforms, get instant alerts, and access every free compliance tool. Or try our Meta Rejection Predictor first.

Create Free Account

Report Keywords — Run AI Compliance Audit

#Brand Safety#Content Moderation#DSA#Enforcement Data#Self-Harm#Violence#Cyber Violence#Adjacency Risk#Advertisers#Agencies#Compliance Guide 2026

Share This Report

Tweet Share

FIFA World Cup 2026 Advertising Compliance: Ambush Marketing, IP Rules and Platform Enforcement

The FIFA World Cup 2026 draws non-sponsor brands to event marketing — and to FIFA's trademark rules, ambush-marketing limits, and the counterfeit and IP policies that platforms enforce on ads.

AI-Generated Ad Content Disclosure Compliance 2026 — Google Ads AI Label, Deepfake Ban & Synthetic Media Rules Across Platforms

Google Ads now requires an AI Generated label on every ad featuring synthetic media and bans deepfakes of real people. Here is the cross-platform AI disclosure compliance framework for 2026.

AI-Generated Ads Legal Compliance 2026 — New York Synthetic Performer Law, California AI Transparency Act & EU AI Act Advertiser Guide

Three major AI advertising laws take effect in summer 2026: New York's synthetic performer disclosure law (June 9), California's AI Transparency Act (August 2), and the EU AI Act. Here's how advertisers must audit their AI ad pipelines to avoid multi-million dollar fines.

Platforms

Industries

Knowledge

Blog

Harmful-Content Enforcement Surged in 2026: A Brand-Safety Playbook for Advertisers