In today’s Digest, we discuss how YouTube may be billing advertisers against violent and pirated content , Google’s hidden search terms costing advertisers millions, and Law360 staff pushing back on mandatory AI bias tools.

Adalytics Reports YouTube Billing Advertisers for Ads on Pirated/Violent Content

YouTube is once again under fire as research from Adalytics suggests advertisers are being billed, and not refunded, for impressions against channels which violate platform standards including copyright infringement and promotion of hate speech. A recent example of the potential scale of the issue included pirated versions of Lilo & Stitch, which racked up over 200,000 views in the days following the film’s release on the platform, potentially costing Disney millions of dollars in additional sales. 

According to new research from Adalytics, the broader challenge of copyright enforcement on the platform reveals how advertisers may be inadvertently funding illicit content. The findings revive long-standing concerns about YouTube’s role in enabling copyright violations and raise new questions around the platform’s content moderation practices.

A YouTube spokesperson, Jack Malon, confirmed that  YouTube does not analyse the roughly 10% of videos removed due to copyright infringement claims, nor does it track how many of those videos may be newly released, full-length movies.

Google's Hidden Search Terms Cost Advertisers Millions

Google maintains that it hides certain search terms to protect user privacy, not to obscure performance data. But fresh analysis of USD$19.6m (£15.4m) in ad spend suggests otherwise, revealing that nearly every dollar spent on Google Ads incurs a hidden efficiency cost tied to restricted search term visibility.

According to a report by Taikun, data collected from 933 campaigns (779 search, 154 shopping) shows that Google Ads performance is materially worse for hidden queries than for those visible to advertisers. In the dataset, a USD$1m (£790,000) ad spend yielded USD$4.75m (£3.75m) in revenue, a 4.75x blended ROAS. However, without the performance gap between visible and hidden search terms, that figure could have exceeded USD$5.6M (£4.42m). 

Another way that Google scrapes profit from advertisers is through the heavy use of hidden search terms. Until September 2020, Google shared almost all search query data. In 2020, Google decided that it would reduce the percentage of search query data that advertisers were allowed to see for “privacy reasons.” I simply do not believe that 28% of all searches coming into Google Ads need to be hidden for privacy reasons.

Law360 Staff Push Back on Mandatory AI Bias Tool

Law360, the legal news outlet, has mandated the use of an in-house AI “bias” detection tool for all editorial content, sparking backlash from its unionised newsroom. The tool analyses drafts for perceived bias and flags lines for revision to ensure a more “neutral voice.” While previously optional, use of the technology became compulsory on 15 May. Reporters are now required to run stories through the tool before publication, alongside other AI-driven functions including headline drafting, tagging, and editing.

The policy shift has prompted strong opposition from the Law360 Union, which represents over 200 editorial staff within the 350-person organisation. In a petition submitted to management on 17 June, union chair Hailey Konnath demanded the tool be made voluntary, arguing that its enforcement undermines journalistic autonomy.

“As journalists, we should be trusted to select our own tools of the trade,” the petition read, calling the AI mandate “unproven technology” imposed “against our will.” The letter was signed by over 90% of union members.