RSF criticizes Google headline rewrites

Reporters Without Borders said Google is claiming an editorial right it does not have by rewriting news headlines in Search with artificial intelligence. (rsf.org) The press-freedom group said the test changes text “written and approved by journalists” and asked Google to end it. RSF said Google confirmed the experiment when the group asked about it. (rsf.org) The dispute traces back to March 20, when Search Engine Land reported that Google had confirmed to The Verge it was testing AI-generated titles in traditional Search results. Google described the rollout as “small” and “narrow,” not approved for broader launch. (searchengineland.com) Search Engine Land said the test affects news sites but is not limited to them, and Google said the goal was to better match titles to user queries. The company also said any future launch might not use generative AI. (searchengineland.com) RSF argued this goes beyond Google’s older practice of automatically generating title links from page elements and references on the web. The group said the new examples show “complete shifts in editorial framing,” not routine truncation. (rsf.org) (searchengineland.com) One example became the centerpiece of the backlash. A Verge headline that originally said an AI cheating tool did not help the reporter cheat was shortened in Google Search to a bare keyword phrase about a “cheat on everything” tool. (rsf.org) (searchengineland.com) Google has said for years that title links in Search are generated automatically from signals such as page titles, headings, anchor text, and structured data. The new criticism is that AI rewriting can alter tone and meaning, not just presentation. (searchengineland.com) The fight is landing as Brazil’s antitrust regulator, CADE, moved on April 23 to deepen a long-running case into Google’s use of journalistic content. Reuters reported that CADE sent the matter back for formal administrative proceedings, citing the evolution of Google’s conduct since 2019. (usnews.com) Brazil’s regulator said the expanded inquiry will examine AI-generated summaries in search interfaces and whether Google extracts value from third-party content without proportional compensation. Valor International reported the CADE vote was unanimous on the AI portion, with five votes. (usnews.com) (valorinternational.globo.com) Google told Reuters it believes CADE’s decision reflects a “misunderstanding” of how its products work and said it would keep engaging with regulators. That response now sits alongside a separate complaint from RSF that Google is no longer just ranking journalism, but rewriting how it is presented. (usnews.com) (rsf.org)