By Fiona Salmon, Managing Director at Mantis
Content moderation has become one of the most heated topics in digital media. Since 2025, several steps have been taken in the US in line with the Trump administration’s views on free speech, including relaxing fact-checking processes for Facebook and YouTube, and the recent decision to ban European figures linked to web media monitoring from entering the country. Over the same period, enforcement of the UK Online Safety Act started requiring digital platforms to actively prevent illegal content, while Australian regulators banned social access for under-16s entirely.
The differing approaches to regulation of digital content have created a divided landscape globally for brands looking to operate across multiple markets. Depending on where you are, the definitions of what is considered suitable content can vary greatly. Amid this division, it may seem as though discussing the need for enhanced brand safety has become too contentious. Yet the risk of placing advertising beside unsuitable content is growing in tandem with tensions. This is exacerbated by the increasing amount of AI-generated synthetic news, manipulated media, and convincingly false narratives entering feeds and exchanges.
Brands are understandably feeling uncertain about where their campaigns should appear, with some travel companies already requesting a complete block on content including ‘ICE’, and more likely to follow. As political messaging increasingly crosses over into content covering key social, sporting, and entertainment events in 2026, as seen at the recent Grammy Awards, brands need to have confidence that marketplace inventory is suitable and safe.
What was previously a niche issue regarding compliance for brand suitability is now a commercial risk management issue for brands, and common approaches for tackling it are not good enough.
The weakness in retrospective systems
Much of the industry remains dependent on retrospective methods to manage unacceptable content. Post-bid monitoring, verification tags, and exclusion reporting are used after the impression has occurred in the advertising supply chain. Should it then be determined that there was problematic adjacency to the ad, adjustments will be made using reporting, credits, or updated block lists.
The flaw in these retroactive systems is that the association has been made, the impression has been counted, and the brand risk has already developed.
There is also the cost associated with allowing non-compliant inventory to circulate through exchanges prior to filtering it out. This creates friction and waste for buyers. Buyers pay for the costs of infrastructure, verification, and reconciliations associated with preventing non-compliant inventory from entering the supply chain in the first instance. These inefficiencies compound as the scale increases.
Another level of distortion occurs due to blunt keyword blocklist systems. Industry analysis has demonstrated that automated filters often pull in legitimate journalism alongside genuinely harmful content, as they do not understand the context in which they exist. In doing so, these systems deprive credible publishers of their deserved revenue stream, and do little to eliminate the root cause of the problem.
Retroactive systems were designed to manage risk, and are not fit to prevent risk at its inception.
Limiting the flow at the source
To create a healthier digital ecosystem, intervention must occur early in the pipeline. The best place to intervene is not at the moment of bid, but at the moment content is created, classified, and distributed.
Publishers and platforms decide what goes into recommendation engines, what is syndicated, and what is eligible for monetization. Decisions made at this stage define the inventory that enters the marketplaces. Once the content is widespread, controlling it is much more difficult and costly.
Today, advances in contextual AI enable intervention at the source of content to be more feasible than it was just a couple of years ago. Contextual AI systems can assess the sentiment of content, the framing of the narrative, and multimodal signals such as those found in text, video, and audio. They can differentiate between content that reports responsibly on sensitive topics and content that is inflammatory, manipulative, and/or intentionally misleading.
By classifying content at the source, the volume of unsuitable content entering the commercial ecosystem initially decreases, and quality journalism is protected from being caught in overly cautious filters. Precise classification at the source enables greater monetization downstream.
This is not an attempt to silence debate or restrict views. Rather, it is an effort to ensure that only content that meets clearly defined standards for advertising eligibility can scale.
Shared responsibility and measurable results
Responsibility does not fall solely to any single segment of the supply chain. Advertisers, publishers, platforms, and technology providers all have roles to play in managing risk.
For advertisers, the process starts with scrutiny. What are the safeguards in the chain? Are controls applied retrospectively, after impressions are served, or are they integrated at the beginning of the content classification and distribution? Are the frameworks used to measure brand suitability taking into consideration the hidden cost of unnecessary filtering and post-bid corrections?
For media owners and platforms, governance and transparency are key. Clear editorial standards, investment in intelligent classification, and visible accountability mechanisms provide confidence to buyers navigating divergent regulatory expectations.
Prevention also provides a measurement advantage. When harmful inventory is filtered at the source, performance data is cleaner, and the ability to measure impact is improved. Buyers can evaluate impact without the noise introduced by misaligned placements or emergency exclusions.
Amid fragmented regulations and an AI-accelerated content environment, retrospective damage control is no longer enough. The industry cannot continue to use cleanup mechanisms after impressions have circulated. Preventing unsuitable content from entering monetizable channels protects brand value, eliminates supply chain waste, and promotes the sustainability of legitimate publishers.
