On YouTube, over 360 hours of content are being uploaded by users every minute. And when it comes to short-form videos like Reels, some 3.5 billion videos are reshared daily across Meta’s platforms. On TikTok, more than 23 million videos are posted every day.
As the number of videos on social platforms continue to grow exponentially, authentic, human-created content is becoming rarer. By next year, some 90% of content online is expected to be generated by AI. For brands, it means preparing for a digital advertising landscape that's more cluttered than ever, making it more difficult to cut through the noise, reach the right audiences, and avoid wasted ad spend.
“This influx will mean a huge increase in video content, much of which advertisers may not wish to be associated with. The need to filter such volumes will be immense, requiring significant technology to process and filter every video,” Marc Grabowski, COO at Integral Ad Science (IAS), told Campaign Asia on a recent visit to Singapore. “Every second of a video could include brand-unsuitable content.”
Placing ads on the right digital platforms has become a massive challenge for advertisers. Adtech companies like IAS are striving to make media buying more effective in combatting ad waste by developing enhanced technology that’s capable of hyper-analysing data and content at scale.
IAS’ ad verification technology now reviews the equivalent of 50 years of video daily, a substantial increase from just two years' worth of similar content in 2023. IAS has engaged this model—Total Media Quality (TMQ)—for major brands like Kimberly-Clark, using AI to sift through every frame of video content to avoid unwanted placements across some 180 countries. The company processes some 280 billion transactions for advertisers daily.
“In terms of key findings, some numbers illustrate its impact: users of our social optimisation can reduce ad waste by 48%, essentially cutting waste in half. Fail rates, which represent placements that are not brand suitable, are reduced by 65% when TMQ is used alongside social optimisation,” explained Grabowski, adding that brand safety—preventing the exposure of ads to universally negative environments—is table stakes.
With the uptick in AI-generated content, ad fraud has also increased, with losses expected to balloon to some $41.4 billion this year. Over 70% of the global ad spend is invested in mobile platforms—where much of social media content is viewed—and 15% of that is being wasted on fraud, with bots and click fraud causing a major drain on digital traffic.
“If advertisers are not looking out for fraud or have no fraud mitigation strategy, they could face up to 19% fraud rates, which can be highly damaging. All learnings feed back into our models, allowing brands to optimise against emerging types of fraud as they are identified,” noted Laura Quigley, senior vice president, APAC at IAS.
“Ultimately, the recommendation would be to have an always-on verification strategy. There will always be peaks and troughs—especially during moments such as elections—but keeping verification active is key, as the solutions to mitigate risks remain available.”
Leveraging AI to ensure brand suitability
Mitigating wavering consumer trust—a result of AI threats on digital platforms such as misinformation, deepfakes, and off-brand content—requires a sophisticated approach to effectively manage campaign integrity. This means focusing on brand suitability to ensures ads appear in contexts that align with specific brand values and engage with the right audience.
“Our research shows that when ads appear next to irrelevant or unsafe content, consumer trust declines, with 70% of consumers expressing distrust towards brands in such scenarios. These are the two key challenge areas: maintaining consumer trust and brand confidence, which are both essential for ongoing investment in these platforms,” Quigley added.
Grabowski observed that within the ad verification process, there’s a shift from focusing solely on brand safety to a more wholistic approach with brand suitability. This approach leverages strong partnerships between adtech companies and brands, so advertisers can optimise supply paths and eliminate inefficiencies.
With its Multimedia Understanding Model (MUM), IAS leverages generative AI to extract audio, images, videos, and texts to analyse and scan content in various contexts, leading to greater transparency for advertisers and maximising programmatic buys with enhanced pre-bid targeting. The model accelerates the ad verification process by 29 times and improves its precision by 45%, providing more accurate validation that results in improved ad performance across various media environments.
Meanwhile, given all the new AI content flooding our feeds, attention continues to stand as a critical metric in evaluating ad performance amid the clutter. To gauge brand suitability for advertisers, IAS looks closely at four key factors: exposure assesses the reach of ads per placement; prevalence reflects the ad’s size and prominence; impact gauges user interaction levels; and clutter considers the density of competing content surrounding ads.
By scrutinising these factors, advertisers can better engage consumers in crowded digital environments. Grabowksi added: “Our partners see our technology not only a layer of protection but a performance driver in their media investments. This framework helps advertisers maximise effectiveness."
