In today’s Digest, we discuss critics arguing that Google’s new Search Console feature masks the decline in AI-driven traffic, Alibaba’s Qwen becoming the world’s most downloaded open-weight AI model, surpassing Meta and Google, and an IAB report finding that Australian agencies are pushing for greater transparency in programmatic DOOH.

Google's new Search Console feature 

Google recently updated Search Console to let brands track how their social media and video content performs within search results, a move many marketers have welcomed as a gesture of goodwill toward companies investing in platforms like TikTok, YouTube, and X. But according to one industry perspective, the update looks quite different once you move past the official framing: rather than a genuine reward, it functions as a strategic shield, one designed to obscure declining website traffic caused by Google's AI-driven search features while simultaneously pulling creators further into training the company's AI models. 

The concern centers on Google's growing reliance on generative search experiences and AI summaries, which answer user queries directly on the search results page rather than sending users to external sites, effectively keeping traffic within Google's own ecosystem instead of routing it to the websites that produced the underlying content. By introducing social media visibility tracking inside Search Console, the argument goes, Google is attempting to redefine what counts as winning: even if a company's website traffic drops by a third, Google can now point to strong impression counts on TikTok videos or other social content as evidence of continued success, encouraging brands to feel reassured despite an underlying shift where fewer users are actually clicking through to their sites.

Qwen overtakes Meta 

Alibaba's open-weight AI models have racked up more than 3 billion downloads globally over the past six months, overtaking Meta, Google, and other domestic Chinese competitors to claim the top spot worldwide. According to Alibaba, its Qwen family of models has open-sourced more than 460 individual models, spawning an ecosystem of over 300,000 derivative projects. 

By comparison, Google logged 418 million downloads and Meta reached 227 million in 2026, according to a "state of open models" report published 14th August by Hugging Face, a widely used open-source AI hub. Because open models can be freely downloaded, customised, and built into new AI products, download and derivative-model counts have become a meaningful signal of which technologies developers are choosing to adopt, making these figures one useful gauge of influence in the broader US-China AI competition.

IAB report flags transparency gaps in Australian DOOH 

Programmatic digital out-of-home advertising is becoming an increasingly established fixture in Australian media planning, with three in four agencies expecting to increase their investment in the channel over the next year, according to the IAB Australia Programmatic Digital OOH State of the Nation 2026 Report. The report identified several areas where buyers are seeking greater transparency, including inventory quality, impression delivery, and reporting on campaign pacing, spend allocation, supply path costs, screen availability, win rates, and proof of delivery. Agencies are also calling for more consistent measurement standards across the industry, including standardised approaches to audience reach, frequency metrics, and impression counting across different vendors.

The report flagged proof of business outcomes and ROI as a critical priority going forward, with the industry increasingly expected to move beyond simple impression and screen counts toward demonstrating tangible business impact.