Classifieds

Google’s AI Model Gemini Spurs Fresh Concerns Among Publishers

If you block the crawlers, then you risk the possibility of blinding Google to your content, and then no one sees it in the first place. Damned if you do, damned if you don’t.

Publisher exec

In September, Google let publishers opt out of having their data used to train Google’s AI models such as Bard

Publishers are also vexed over the level of attribution that AI tools provide to the sources of information, often derived from expensive journalistic efforts. Complicating matters is the legal concept of “fair use,” which permits the use of copyrighted content without explicit permission.

But what makes Google a bigger threat, compared with rivals such as OpenAI, is that the tech giant is a primary gateway to the internet, making its application uniquely concerning, said a content strategy leader at a major publisher, speaking anonymously.

“As Google roadmaps products such as Bard and SGE [search gen AI] to play an increasingly central role in general search and content discovery, a publisher’s SEO [search engine optimization] efforts will increasingly work against them,” the exec said. “If you block the crawlers, then you risk the possibility of blinding Google to your content, and then no one sees it in the first place. Damned if you do, damned if you don’t.”

An increase in official deals

The industry expects a notable increase in official agreements between Google and publishers, similar to that of the Associated Press and OpenAI, especially in scenarios with high-quality data, such as that held by news publishers, said Katie Gardner, partner at law firm Gunderson Dettmer.

Data licensing agreements between content providers and content distributors hinge on several parameters, such as the value and timeliness of the data and the potential revenue offered to publishers.

“The data owner will want to maintain as much control as they can and want to preserve optionality in how to monetize that data, whether it’s licensing to other foundational models, or building their internal models,” Gardner said. “Ultimately, it just comes down to the dollars.”

[ad_2]
Source link

Previous page 1 2

Related Articles

Back to top button