Since August 2, 2026, a new set of obligations under the European AI Act applies. Among them, Article 50 sets a simple rule in principle: any content, text, image or video, generated or modified by artificial intelligence, must be disclosed as such when shared publicly. A company that posts AI-generated visuals or text on its social media is directly concerned.
This is above all a transparency rule. The highest sanctions provided for by the AI Act (up to €35 million or 7% of global revenue for the most serious breaches) primarily target high-risk systems and foundation model providers, two categories that rarely apply to a company using an AI community management tool for its own social media. That does not make it a secondary point: the transparency obligation applies from the very first post.
An AI tool designed after the fact to "become compliant" does not approach the topic the same way as a tool built with that constraint from the start. Isalis was designed and hosted for Europe from day one, even before this kind of compliance was widely required: that choice slowed its early development, but today it avoids having to rebuild the tool to meet obligations that are phasing in progressively.
Isalis also chose, from the outset, never to publish content automatically without prior human review. That review step is precisely the right moment to check whether an AI-generated content disclosure needs to be added, when Article 50 requires it: the check is not bolted on afterward, it relies on a control already built into how the tool works.
Two independent services, activated according to your actual needs.
Learns your brand's visual identity and tone, then creates, schedules and publishes visuals, videos and text on your social media, with human review before anything goes live.
To go beyond a single tool: an inventory of every AI system used in the company, including undeclared usage, risk classification and a compliance register.
These two services are sold separately and do not depend on one another: a company can use Isalis without commissioning a diagnostic, and vice versa.
It requires that any content (text, image or video) generated or modified by artificial intelligence be disclosed when shared publicly. It is a transparency obligation, not a pre-authorization regime.
Yes, in principle: Article 50's transparency obligation covers any public distribution of AI-generated content. The AI Act's most stringent obligations concern high-risk systems and foundation model providers, two different categories that do not apply to community management use.
Isalis is not meant to replace a legal analysis of your situation. What is already in place: no post goes out without human review, which is precisely the moment when that disclosure can be checked and added if it is required for the content in question.
Isalis's European hosting and design primarily address GDPR, not directly Article 50 of the AI Act, which concerns content transparency rather than data location. The two topics are connected by the same logic: building the tool with European rules in mind from the design stage, rather than adapting to them afterward.
That is exactly the kind of situation Sundatalab's AI Act diagnostic covers: an inventory of the AI tools actually used in the company, including undeclared ones, risk classification, and a ready-to-use compliance register.
Features, use cases, pricing and a trial: it's all on isalis.ai. For a compliance need broader than this one tool, the AI Act diagnostic remains an independent service.