Brand safety, making sure ads do not appear alongside content that could damage a brand, has always been part of digital advertising. For years it was largely handled with blocklists of sites and keywords, and a set of shared industry standards.
Generative AI has made the problem harder. Text, images, audio and video can now be produced quickly and cheaply, at enormous volume. Some of this content is useful. Much of it is low-quality material designed mainly to attract ad revenue, and some is misleading or deliberately deceptive, including deepfakes (realistic fake videos or audio of real people). At the same time, the industry has lost a shared reference point: the Global Alliance for Responsible Media, an industry body that coordinated brand-safety standards, closed in 2024.
For advertisers, this means brand safety needs more attention, and a more thoughtful approach, than it once did.
The new risks
Three risks have grown most clearly.
| Made-for-advertising sites. | Websites built mainly to capture ad spend, filled with low-value or automatically generated content and heavy with ads, can absorb budget while delivering little real attention. |
|---|---|
| Misleading and synthetic content. | Ads can appear alongside fabricated news, false claims or manipulated media, which risks associating the brand with misinformation. |
| Impersonation. | Generative tools make it easier to create fake ads or accounts that use a brand's name, logo or even the likeness of its spokespeople to deceive customers. |
Why blocklists are not enough
Keyword and site blocklists are still useful, but they have limits.
They are blunt. Blocking broad words like "crisis" or "health" can keep ads away from large amounts of legitimate, high-quality journalism, which harms publishers and reduces the brand's reach among valuable audiences.
They are slow. New sites and content appear constantly, and a list is always behind.
They look at words, not meaning. Much of the new risk lies in whether content is genuine, accurate and made for people, which a keyword cannot judge.
A more balanced approach
Effective brand safety today combines several practices.
Focus on quality, not only avoidance.
Rather than only listing what to avoid, define what good looks like. Inclusion lists of trusted publishers and channels, and buying through curated marketplaces, keep more spend in environments that are known to be safe and effective.
Use contextual tools that read meaning.
Newer contextual technology analyses the actual content of a page or video, not just its keywords. This allows more precise decisions, avoiding genuinely harmful content without blocking entire topics.
Demand transparency.
Ask partners for clear reporting on exactly where ads appeared. Placement-level reporting makes it possible to spot and remove poor environments quickly.
Watch for impersonation.
Monitor platforms for fake ads and accounts that use your brand, and use each platform's reporting tools to have them removed. Clear, consistent official channels help customers recognise what is real.
Support content provenance.
Industry standards such as C2PA (Coalition for Content Provenance and Authenticity), which attach verifiable information about how a piece of content was created and edited, are becoming more widely adopted. Supporting and using them helps build a healthier environment over time.
Balancing safety and reach
Brand safety is not only about risk. Overly strict controls can remove ads from valuable news and cultural content, cutting reach and funding for quality publishers. The goal is suitability: placing ads in environments that match the brand's values and audience, not simply avoiding every possible issue.
Protecting the brand, funding real attention
Generative AI has increased both the amount of content and the uncertainty about what is genuine. Brand safety now calls for a more active approach: favouring trusted environments, using tools that understand context, demanding transparency and watching for misuse of the brand itself. Done well, it protects the brand and puts the budget where real audiences are paying attention.
