The digital landscape is transforming under the combined effect of new regulations, the massive integration of artificial intelligence, and security constraints that are reshaping companies’ priorities. Understanding these changes requires distinguishing between what falls under the already applicable legal framework and what remains in the realm of technological trends.
European AI Act: a regulatory timeline changing the governance of digital tools
The European regulation on artificial intelligence (AI Act) is not just a theoretical text. Its initial provisions, concerning the control of AI and prohibited practices, have been in effect since February 2025. The obligations related to general-purpose AI models came into force on August 2, 2025.
This phased deployment requires organizations to undertake concrete mapping work: identifying which AI systems are used internally, assessing their risk level according to the regulation’s classification, and documenting the associated decision-making processes. Training teams is part of the obligations, not just the technical compliance of the tools.
For digital marketing or e-commerce professionals, the direct consequence affects personalization tools, customer scoring, and content generation. Each tool integrating an AI model must now be subject to governance oversight. Following news on Been allows for tracking these regulatory developments as they are implemented.
Digital accessibility: the European directive 2019/882 applicable since June 2025
Since June 28, 2025, the European Accessibility Act applies to several categories of services, including online commerce and consumer banking services. The scope depends on the service in question; it is not a uniform obligation for all websites.
In practice, companies selling online must verify the compliance of their purchasing process with the accessibility criteria defined by the directive. This includes keyboard navigation, text contrast, compatibility with screen readers, and the semantic structuring of HTML content.
What the directive concretely covers
- E-commerce websites, including payment and order tracking interfaces
- Digital banking services for consumers (mobile apps, online customer spaces)
- Certain transport platforms offering online booking
Accessibility is no longer an optional improvement area. For the affected services, non-compliance with the directive exposes organizations to sanctions defined by each member state. Web teams and front-end developers are the first impacted in adapting existing interfaces.
Cybersecurity and AI: a rapidly expanding attack surface
The integration of AI into digital environments creates new vulnerabilities. The European agency ENISA has documented the rise of threats related to environments integrating artificial intelligence, particularly attacks targeting software supply chains and language models exposed via APIs.
This progression directly affects highly interactive user sites: e-commerce, contact forms, member spaces. Detection solutions must evolve at the pace of attack techniques, which now exploit generative models themselves to bypass traditional protections.
Three areas of vigilance for companies
Cybersecurity in a context of widespread AI is no longer limited to firewalls and antivirus. Organizations must prioritize monitoring:
- The security of APIs connected to AI models, which represent a prime entry point for prompt injection attacks
- The reliability of training data used by internal tools, as a compromised dataset skews downstream automated decisions
- The management of access to generative AI tools by employees, to prevent leaks of sensitive data in prompts sent to third-party services
These security issues do not only concern IT departments. Marketing teams using content generation or predictive analysis tools handle customer data, the protection of which is also their responsibility.
Generative search and web visibility: what changes for content
Google is gradually rolling out updates to make sources more clickable in its AI-generated responses. These adjustments come as content publishers are noticing a drop in organic traffic related to AI Overviews, which provide answers directly on the results page.
Natural referencing is not disappearing, but content must now be structured to be picked up by generative engines. This means clear data, rigorous semantic markup, and identifiable factual answers within the text. Long and vague articles are losing ground to content that precisely answers an identifiable question.
The current period is characterized by the simultaneity of these changes. AI regulation, mandatory accessibility, cybersecurity threats amplified by generative models, and a rethinking of visibility logic on search engines: these four areas are not distant trends. Each of them is already producing measurable effects on budgets, required skills, and the tool choices of digital teams.
