A number of Fortune 500 brands–including Microsoft, Meta, Disney, and more–are having their ads served next to harmful content online including pornography and pages with racist content, according to a report from Adalytics.
Of these ads, some included a Javascript from vendors including software company DoubleVerify and digital ad tech and verification company Integral Ad Science (IAS), both of which use AI-powered brand safety tools.
The report questions the effectiveness of current brand safety technologies, especially ones that offer protection at a URL and page level, and marketers as a result are asking for full URL transparency from demand-side platforms and the need for these tools to be updated to provide a higher level of security.
Since the report was published, the IAS released a statement which claimed that the report from Adalytics 'inaccurately represented' its technology and that it failed to take into consideration measurement (post-bid) and optimisation (pre-bid) settings. It also said that the report conflates IAS code used for publisher solutions and code used for advertiser solutions.
The statement said: “The code analysis conducted by Adalytics is not representative of the full suite of tools IAS provides to advertisers and publishers with multiple layers of brand safety and suitability measurement and protection.”
“The examples shared only represent one part of our brand safety and suitability solution. Another flaw in the report shows IAS tags on a page that do not correspond to a final decision for the creative being displayed.”
DoubleVerify also published a blog post in response to the report, claiming that the results in the report are 'entirely manufactured' and that no customer has expressed concerns about the accuracy of its content categories. It also claimed that other reports by Adalytics, including the MFA and Forbes ones, are also inaccurate.
In response to the report, Arielle Garcia, director of intelligence at Check My Ads, posted on LinkedIn that advertisers should demand page URL level report and more clarity on how the tools work. She also said publishers should demand more transparency, as publishers with 'benign' content are being labelled as high-risk and risk being demonetised.
She said: “Brand safety vendors' black-box tech, and 'set-it-and-forget-it' promises have yet again failed advertisers and publishers.
“These false promises of AI-powered EZ-buttons are irresponsible and harmful. At Check My Ads, we've been flagging their shoddy content classification since we tested out IAS’ quickly-removed “Context Control” demo tool back in 2020.”
PMW has reached out for a comment from IAS and DoubleVerify.
The story was originally published on Campaign's sister publication Performance Marketing World.