A complaint built on a documented enforcement history
The Alliance de la Presse d'Information Générale (APIG), representing nearly 300 French dailies including Le Monde, Le Figaro, Les Échos, Ouest-France and Le Parisien, filed a complaint against Google with France's Autorité de la concurrence on August 11-12, 2026. The claim: Google's AI Overviews and AI Mode use publisher content to generate answers directly in search results without prior authorisation or dedicated remuneration, breaching commitments Google made to the regulator in 2022.
This is not APIG's first swing, and it isn't starting from zero. Google was fined €500 million in July 2021 for failing to comply with 2020 interim injunctions to negotiate with publishers over neighbouring rights. It then accepted a five-year (renewable) commitments package in June 2022 — Decision 22-D-13 — requiring good-faith negotiation on "transparent, objective and non-discriminatory criteria," disclosure of impression and revenue data, and arbitration if talks stalled. In March 2024, the Autorité fined Google another €250 million — Decision 24-D-03 — for breaching four of those seven commitments, including training Bard on publisher content without notification and tying AI-feature opt-outs to loss of other protections. APIG's new complaint is effectively an argument that Google is repeating the 24-D-03 pattern with AI Overviews.
The publishers' case, stated fairly
The strongest version of APIG's argument doesn't require sympathy for legacy media — it requires taking the 2022 commitments at face value. Google agreed to negotiate compensation for new uses of protected content. AI Overviews synthesize publisher reporting into a standalone answer at the top of the results page; the reader's information need is met without a click. Arcom, France's broadcast and digital communications regulator, has estimated AI-generated summaries account for a 33-38% drop in referral traffic to publisher sites in markets where the features are live, a figure APIG has folded directly into its filing, per Euronews. A separate Ahrefs study of 300,000 keywords found top-ranking-result click-through rates fell 58% for AI Overview searches between December 2023 and December 2025, according to MediaNama. If a platform's dominant search product structurally suppresses the traffic that funds the journalism it depends on, and does so via a product change publishers can't meaningfully opt out of without disappearing from search entirely, that is exactly the kind of unilateral, unremunerated "new use" the 2022 settlement was written to prevent. Given Google's actual track record of non-compliance, APIG's skepticism that this will resolve without regulatory pressure is earned, not paranoid.
Where the case gets weaker
The steelman thins out once it moves from "Google should pay something" to "the Autorité should treat AI Overviews as a simple repeat of the Bard breach." AI Overviews are not a separate product bolted onto search — they are, functionally, an evolution of the answer boxes and knowledge panels Google has served for over a decade, changes publishers never had a standalone right to block. Google has publicly said it has "reassessed its system for paying the press" to account for AI-feature display and maintains publishers can choose not to appear in AI features, per Euronews — a claim the Autorité, not this publication, is best placed to test against the actual mechanics of that opt-out.
More importantly, France's complaint is not happening in isolation. The European Commission opened its own antitrust investigation into Google's use of publisher and YouTube content for AI Overviews and AI Mode on December 9, 2025, examining nearly identical questions of compensation and opt-out coercion, according to Foxglove. Running a national neighbouring-rights case in parallel with a Brussels-level DMA-adjacent probe over the same conduct risks exactly the kind of fragmented, double-jeopardy enforcement that makes compliance unpredictable for every AI product built on web content — not just Google's. A platform facing simultaneous, differently-scoped investigations in Paris and Brussels has less certainty about what compliance even looks like, which is bad for publishers hoping for a fast remedy and worse for the next AI search entrant trying to build a compliant product from scratch.
The proportionate path
The 2022 commitments regime — negotiated remuneration, transparent metrics, arbitration for smaller publishers — is a defensible model precisely because it doesn't require blocking a search feature outright; it requires paying for it. That's the outcome APIG says it actually wants: a return to the negotiating table, not an injunction against AI Overviews in France, per VideoWeek. Regulators should hold Google to the compensation framework it already agreed to rather than let a live enforcement mechanism sit idle while a slower EU-level probe grinds through 2026 and 2027. The Autorité's own history here — three enforcement actions against one company on one legal theory in five years — argues that a negotiated remuneration fix is achievable without freezing the underlying product. The real test is whether France pursues that fix through the existing 22-D-13 machinery, or lets this become the opening motion in a multi-jurisdictional standoff that outlives the traffic numbers it's meant to address.