Last week, we saw another raft of artificial intelligence Ads of Google and Openai. The latter brought (not) the voice of Scarlet Johansson to a vocal assistant in real time, seated on his new model Cat 4o (‘ninety). The first with a disappointing conference for the developers in which he revealed – perhaps for lack of something more interesting to say – a faster deployment of the “ ia ” sewn (or technology formerly known under the name of `Generative Research Experience ”). All English-based users based in the United States will now see this functionality.
AI’s overviews really count for publishers. This development presents to consumers an entirely formed response and in natural language to a research request. It is above blue links on the search engine results page and is expressly designed to cancel the need to visit another website.
As google promotional video says: “We are doing the job, so you don’t have to do so”. The marketing film continues to show us how we can ask for “adapted to children” activities in Dallas and go to the “Hopdoggy Burger Bar” followed by the Dallas zoo without second thought.
(The image below compares how the term niche search “conversion rate of publishers” results in an article rewritten in summarizing the content of the publisher on Google’s search engine in the United States (in left) while in the United Kingdom, it translates into a series of links to publishers, including the Gazette press.)
“ Grand Bargain ” from Google with publishers, is now broken
The problem here is that AI’s overviews only work because content creators, including publishers, gave Google permission to crawl their sites so that they are indexed to research. This is a violation of the great affair already a thousand in place between Google and the publishers.
In the case of conventional research, negotiations go something like that. We give Google access to our intellectual property, allowing the technology giant to create its basic product, and in exchange, Google offers traffic which can then be monetized by advertising, subscriptions, events, etc.
Put aside the imbalance of negotiation power which raises questions on the question of whether it is a good deal, there is at least an exchange.
The AI glimps interrupt this because Google does not deliver its part of the negotiation: reference traffic. To make things worse, Google does not let content creators withdraw from IA glimps without also withdrawing from its basic research product.
In the Developer notes for research He says to us: “The IA previews offer an overview of a subject or a request based on a variety of sources, including web sources. As such, they are subject to research preview controls. »»
The last torsion of the knife – as if we needed – is that it is now Service announcements Against IA’s overviews.
Just to clearly define this: Google uses access to the editor’s content than points of sale have no choice but to give it – thanks to the goalkeeper position he holds by connecting to a Public – To create a service that directly replaces the use of our product. And then it is monetizing this service by advertising, probably to the detriment of content creators. Any without authorization or payment.
An injustice and an abuse of market power
In addition to being an injustice, it is surely an abuse of market power. The authority of French competition judge Against this driving earlier in the year. The fact that Google has chosen to continue to act in this way (it actually widened the type of condemned behavior) points out the extent to which the people of Mountain View are frightened by the risks that participants in his research monopoly.
Yes, this is Openai. But also, I suspect that the perplexity LLM Rival bout, which targets research more directly. As a holder, Google will know that the regulatory and legal reaction will be more difficult, faster and will have greater reputation impacts. The immediate risks downwards for his core business not to act in this way, he must have calculated, is higher.
So what can publishers do about it?
How publishers can protect themselves from Google AI threatened
I was responsible for modeling the commercial risks halfway through for publishers of British news of the deployment and adoption of AI. Although the full report has not yet been published (look at this space), it provides both a projected impact scale for different categories of new publishers (which can be refined at the level of the individual editor) And strategic clues on how publishers can best protect themselves.
Although the precise optimal response of each publisher will depend on the characteristics of their business model and their content offer, there are three major recommendations that would apply universally:
First, and of course, reduce your dependence on Google. Establish direct relations with your readers. Build communities. Develop your newsletters. Develop your database.
Second, be known to create content that counts. For some research requests, the source of information sought is real consequence. Think if you choose a car or if you get interview advice. AI research will have lower utility in these circumstances, because users will want brands information they know. If your publication is reliable to provide content of this nature, you will have some protection.
Finally, the distinctive character. The usefulness of chatbots to respond to requests on current events depends on their access to real -time information from content providers. For some subjects, there is a rarity of such entries. Can you provide this information? And even better, be known and trusts it?
Then there is the commentary that leads the speech; The Summary of the AI cannot supplant the reading, of the first hand, this columnist around national, local or sectoral debate centers. Be the supplier of this.
The media industry needs a regulatory intervention urgently to stop the driving of Google. In the United Kingdom, this does not seem a short-term perspective (Lucy Frazer announcement From the government’s desire to create a “framework or politics” does not inspire confidence and the next elections make even more out of words).
In the absence of this, the publishers would do well to decline from the research giant and to concentrate with precision on the information needs with high issues of their readers with distinctive content.
E-mail pged@pressgazette.co.uk To underline errors, provide history advice or send a publication letter to our blog “Letters”