Artificial intelligence is increasingly appearing in Google's services, sometimes where its presence seems questionable. As noted by journalists from The Verge, Google has started experimenting with replacing article headlines in Google Discover with generated versions that sometimes do not align with the content of the publication.

In some instances, the rewritten headlines were misleading or false. For example, Discover showed an Ars Technica article with the headline "Steam Machine price revealed", even though the original article only discussed the design of the device and provided no information about Valve's pricing. The Engadget editorial team also observed another format of the experiment where original headlines were accompanied by generated brief summaries.

Google labels these elements as "Generated with AI, which can make mistakes", but this hasn’t stopped users from criticizing the new feature for its inaccuracies and the risk of misinformation.

Google representative Mallory DeLeon informed The Verge that this is a "small UI experiment" for a subset of Discover users. According to her, the goal is to "change the placement of headlines to make it easier for users to familiarize themselves with the topic before moving on to other materials".

Despite this, the experiment raises concerns among online media that have had conflicts with Google over the use of their texts in search and news services. In previous years, publishers have demanded compensation for displaying their content, while Google has responded by restricting access to their materials in search.

Meanwhile, Google continues to integrate AI more actively into its core search functions. Google Search VP Robby Stein announced that the company is testing the combination of AI Mode and AI Overviews on a single mobile search screen, which will make AI suggestions even more visible and impactful in the results.