The web trends of late 2026 are not just a catalog of design novelties or a tech news feed. Three underlying movements are simultaneously reshaping how websites attract traffic, comply with European law, and protect their youngest audiences.
AI Act and transparency of generative systems: the framework applicable since August 2026
The transparency rules set out in Article 50 of the AI Act came into effect on August 2, 2026. Any provider of a generative AI system must clearly inform the user that they are interacting with a machine.
The European regulation 2026/1744, known as the “Digital Omnibus,” has simultaneously postponed the requirements for high-risk systems listed in Annex III to December 2, 2027. As a result, platforms that integrate AI into sensitive processes (recruitment, scoring, automated moderation) operate in a regulatory gray area, without yet being subject to enhanced obligations.
For website publishers, the consequence is tangible. Their chatbots or virtual assistants must display their artificial nature. Upcoming requirements regarding the documentation of the models used should be anticipated now.
We observe that many French websites have still not updated their legal notices or AI policy pages, even though the framework is already in effect.
Traffic decline from Google: the post-classic SEO era for news sites
Finding all the news on EpicBuzz remains possible, but the path from Google has changed. AI Overviews and AI Mode synthesize answers directly in the results page. The traffic decline for publishers that results from this is measurable.
Google has recently adjusted the formatting of its AI responses: links positioned higher, more visible attributions. These changes aim to make sources more clickable.
Retention outside search engines
The Reuters Institute has documented how newsrooms are adapting their strategies in response to what it calls “life after SEO.” Newsletters, proprietary apps, closed communities: these channels are becoming priorities for audience retention. Natural search is no longer the main acquisition lever for news media, even though it has not disappeared.
In France, the ibou Actualités project illustrates this reconfiguration. Designed as an alternative to AI Overviews, it aims to direct traffic back to media sites rather than absorbing it into an aggregated response.
EU KIDS Act: the European proposal for online minor protection
Presented by the European Commission on September 17, 2026, the EU KIDS Act project introduces a specific framework for those under 18 on digital services. The proposal includes:
- Access restrictions to social networks adjusted according to the user’s age, with mandatory verification by platforms
- An increased responsibility for service providers regarding the safety of their interfaces for young audiences
This text remains a proposal submitted to Parliament and the Council. It is not yet in effect. Its adoption timeline remains uncertain, but it extends the Digital Services Act and the GDPR by tightening the provisions applicable to this age group.
For website and application creators, the EU KIDS Act would require rethinking the design of the user experience. Reliable age verification systems without excessive data collection present the main technical challenge, as no European standard has yet been officially validated.
Source visibility in AI responses: a structural problem for the open web
Google has increased the links to publishers in its AI Overviews. This increased visibility does not resolve the underlying issue: the click remains optional when the answer is already displayed.
A link in an AI response does not generate the same traffic as a classic organic click. The user who gets their answer without leaving Google has no reason to visit the source site, even if it is cited. Counting displayed links without measuring actual clicks gives a misleading picture of the situation.
AI bots and bandwidth consumption
According to DataDome, the traffic from malicious bots has increased by 124% in one year. AI training bots, even when not classified as malicious, consume considerable server resources. Several publishers have begun blocking these crawlers via their robots.txt file, risking a loss of visibility in AI responses.
- The robots.txt file is not a technical barrier, only a convention that bots may or may not respect
- The contractual clauses of the terms of use have not yet been tested in court in most European jurisdictions
- The distinction between indexing bot (useful for SEO) and training bot (which extracts content without generating traffic) remains blurred in default configurations
The phrase “open web does not mean open bar,” promoted by cybersecurity actors, summarizes the tension. Publishers must balance exposure and protection of their content, without having stabilized legal tools.
Three decisions will structure the coming months: effective compliance with the AI Act, legislative progress of the EU KIDS Act, and the ability of publishers to maintain qualified traffic in the face of attention capture by AI responses. These issues touch different layers of the web but converge towards the same conclusion: the web page is no longer the default endpoint of the online user experience.
