Five French publisher organisations filed a complaint on Monday against Google's AI before the European Commission. The same day, the New York Times and several American dailies sued the Brave browser, accused of scraping their articles and reselling that data to AI companies. Two lawsuits, two continents, one target: the way AI uses the web without paying for it.
We see the bargain that funded the open web for twenty years coming apart. The bargain was simple: a publisher published, Google indexed, the user clicked, advertising and subscriptions followed. AI-written answers, displayed at the top of results, break that circuit by answering in the user's place, so the user no longer clicks.
The day before, the same big players proposed a signal in the robots.txt file to refuse training; non-binding, it convinced no one. The courts were what remained.
For a French organisation, the signal goes beyond the press. Any activity that lives on search traffic, a media outlet, an online store, a software vendor that draws visitors with its content, will see that traffic shrink as AI answers capture the click.
The way out lies in channels the search engine does not control: a subscriber list, a loyal audience, a community, a brand people seek out by name. The organisations that build this direct relationship today will keep an audience when the search tap is turned off.
The front is technical too. Researchers show this week that a model's tendency to spit out its training data word for word can be reduced by pruning certain parts of the network. The battle over content is being fought at once before the judges and in the architecture of the models.
The open web was built on an exchange: visibility for content. AI takes the content and keeps the visibility. The asset that will matter in six months is the direct relationship you have built with your audience, for the day the search engine stops sending it your way.
The Masteria editorial team.
