À suivreGuide suivant
Chatbots IA, enfants et lois sur la sécurité en ligne
Société
GUIDE DE LA SOCIÉTÉ
The UK Online Safety Act regulates defined services, not every chatbot simply because it generates text.
Ofcom says coverage depends on whether functionality meets statutory definitions such as user-to-user, search or pornography-publishing services, so product features and sharing flows matter.
The Online Safety Act 2023 creates duties for providers of defined online services with a UK connection. Ofcom identifies three chatbot-relevant categories: user-to-user services that let people share content; search services enabling searches across more than one website or database; and services publishing pornographic content, which have age-assurance requirements. A chatbot can be covered by itself or as part of a larger platform. Ofcom describes a chatbot as potentially outside these categories when people interact only with the bot, it does not search multiple sites or databases, and it cannot generate pornography. This is a functional boundary, not a general exemption for products marketed as AI. In February 2026 Ofcom explained it could not investigate creation of illegal images by the standalone Grok chatbot under the Act’s current scope, while it separately examined X’s social platform and handling of content. Ofcom supports government consideration of possible changes, but discussion does not itself amend legislation. For covered services, obligations depend on service type and content. User-to-user and search providers have illegal-content risk-assessment and mitigation duties, along with applicable children’s safety requirements. AI material shared by users on a covered user-to-user service counts as user-generated content, whether created there or elsewhere. Pornography-publishing services must use highly effective age assurance. Ofcom enforces the Act; explanatory pages clarify statutory scope but do not expand it beyond the definitions Parliament enacted.
Les dommages catastrophiques et quotidiens causés par l’IA dépendent tous deux de la personne qui comprend les risques et qui peut agir.
Les connaissances du public et des professionnels déterminent si une politique de sécurité forte est politiquement possible.
Des explications claires réduisent la capture par le battage médiatique, les relations publiques en laboratoire et le théâtre d'éthique vague.
Ofcom continues to study chatbot risks and government has considered whether the framework should change. Research notices and discussion papers show regulator activity, while reforms remain proposals until legislation changes. Providers should assess their live features against the Act and monitor Ofcom codes, guidance and enforcement decisions. If Parliament changes scope, update the service map rather than assuming Ofcom’s current chatbot example remains unchanged. Keep evidence for moderation, age assurance and user-sharing paths. Monitor Ofcom statements and enacted amendments. Record launch dates.
A private one-to-one companion bot that does not search several sites or generate pornography may fall outside Ofcom’s described service categories.
A platform lets users publish custom bots and share conversations, which may make it a user-to-user service.
An AI answer engine searches multiple websites to respond and may meet the search-service definition.
Users post AI-generated abusive images on a covered social service; the posts are treated as user-generated content.
Traiter le risque existentiel comme de la science-fiction alors que les capacités s’accroissent.
Confondre sécurité des produits de surface et alignement sous haute autonomie.
Laisser le public non anglophone et non expert avec uniquement des sources de mauvaise qualité.
Séparez les dommages causés aux produits, leur mauvaise utilisation et les risques de perte de contrôle/désalignement.
Demandez quelles preuves pourraient changer votre point de vue sur les délais et la gravité.
Préférez les sources primaires et les évaluations concrètes aux allégations marketing.
Identifiez une voie d’action : carrière, politique, financement ou compétences – et pas seulement la sensibilisation.
Free newsletter
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
The UK Online Safety Act regulates defined services, not every chatbot simply because it generates text. Ofcom says coverage depends on whether functionality meets statutory definitions such as user-to-user, search or pornography-publishing services, so product features and sharing flows matter.
User-to-user status turns on enabling people to share content, not the model architecture.
Ofcom describes a narrow outside-scope example using those three functional criteria.
Multi-site or multi-database search may meet the search-service definition.
Ofcom says it is regulated like comparable human-generated user content.
Ofcom distinguished standalone creation from platform distribution and duties.
Continuez à apprendre
Plus de guides sélectionnés pour ce sujet
À suivreGuide suivant
Chatbots IA, enfants et lois sur la sécurité en ligne
Société